Scholarly article on topic 'A study to evaluate the effect on Mouth Level Exposure and biomarkers of exposure estimates of cigarette smoke exposure following a forced switch to a lower ISO tar yield cigarette'

A study to evaluate the effect on Mouth Level Exposure and biomarkers of exposure estimates of cigarette smoke exposure following a forced switch to a lower ISO tar yield cigarette Academic research paper on "Health sciences"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{Cigarette / Smoke / Exposure / Nicotine / Biomarkers / Carcinogen / Dose / "Filter analysis"}

Abstract of research paper on Health sciences, author of scientific article — Christopher J. Shepperd, Alison C. Eldridge, Graham Errington, Michael Dixon

Abstract A forced switch to a lower ISO tar yield cigarette was used in a clinical study, conducted in Germany, that compared two methods of estimating exposure to cigarette smoke. Pre- and post-switch estimates of Mouth Level Exposure (MLE) to nicotine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), pyrene and acrolein were obtained by chemical analysis of spent cigarette filters for nicotine content. Similarly, pre- and post-switch estimates of uptake of these smoke constituents were achieved by analysis of corresponding urinary biomarkers of exposure (BoE): total nicotine equivalents; total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL); total 1-hydroxypyrene (1-OHP), and 3-hydroxypropyl-mercapturic acid (3-HPMA), plus the nicotine metabolite cotinine, in plasma and saliva. Three hundred healthy volunteers were recruited comprising 100 smokers of each of 9–10 and 4–6mg ISO tar yield cigarettes and 50 smokers of 1–2mg ISO tar yield cigarettes and 50 non-smokers. Fifty smokers of each of the 9–10 and 4–6mg ISO tar yield cigarettes took part in the switching aspects of this study whilst the remaining smokers formed non-switching control groups who smoked their usual ISO tar yield cigarette throughout the study. After 5days, all subjects were admitted into a clinic where baseline measures of MLE and BoE were obtained. The 10mg switching group was then switched to the 4mg ISO tar yield cigarette and the 4mg ISO tar yield switching group switched to the 1mg cigarette. Subjects returned home for 12days, continuing to smoke the supplied cigarettes before being readmitted into the clinic where samples were collected for MLE and BoE analysis. Changes in daily exposure estimates were determined on a group and individual basis for both methods. The pre- to post-switch directional changes in MLEs and their corresponding BoEs were generally consistent and the MLE/BoE relationship maintained. Switching to a lower yield cigarette generally resulted in reductions in exposure with the resultant exposure level being similar to that seen in regular smokers of the lower yield cigarette.

Academic research paper on topic "A study to evaluate the effect on Mouth Level Exposure and biomarkers of exposure estimates of cigarette smoke exposure following a forced switch to a lower ISO tar yield cigarette"

Contents lists available at ScienceDirect

Regulatory Toxicology and Pharmacology

journal homepage: www.elsevier.com/locate/yrtph

A study to evaluate the effect on Mouth Level Exposure and biomarkers of exposure estimates of cigarette smoke exposure following a forced switch to a lower ISO tar yield cigarette

Christopher J. Shepperda'*, Alison C. Eldridgea, Graham Erringtona, Michael Dixon b

a British American Tobacco, Group Research and Development, Southampton, UK b Dixon Consultancy, Liphook, UK

ARTICLE INFO

Article history:

Available online 6 June 2011

Keywords:

Cigarette

Exposure

Nicotine

Biomarkers

Carcinogen

Filter analysis

ABSTRACT

A forced switch to a lower ISO tar yield cigarette was used in a clinical study, conducted in Germany, that compared two methods of estimating exposure to cigarette smoke. Pre- and post-switch estimates of Mouth Level Exposure (MLE) to nicotine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), pyr-ene and acrolein were obtained by chemical analysis of spent cigarette filters for nicotine content. Similarly, pre- and post-switch estimates of uptake of these smoke constituents were achieved by analysis of corresponding urinary biomarkers of exposure (BoE): total nicotine equivalents; total 4-(methylnitrosa-mino)-1-(3-pyridyl)-1-butanol (NNAL); total 1-hydroxypyrene (1-OHP), and 3-hydroxypropyl-mercaptu-ric acid (3-HPMA), plus the nicotine metabolite cotinine, in plasma and saliva. Three hundred healthy volunteers were recruited comprising 100 smokers of each of 9-10 and 4-6 mg ISO tar yield cigarettes and 50 smokers of 1-2 mg ISO tar yield cigarettes and 50 non-smokers. Fifty smokers of each of the 910 and 4-6 mg ISO tar yield cigarettes took part in the switching aspects of this study whilst the remaining smokers formed non-switching control groups who smoked their usual ISO tar yield cigarette throughout the study. After 5 days, all subjects were admitted into a clinic where baseline measures of MLE and BoE were obtained. The 10 mg switching group was then switched to the 4 mg ISO tar yield cigarette and the 4 mg ISO tar yield switching group switched to the 1 mg cigarette. Subjects returned home for 12 days, continuing to smoke the supplied cigarettes before being readmitted into the clinic where samples were collected for MLE and BoE analysis. Changes in daily exposure estimates were determined on a group and individual basis for both methods. The pre- to post-switch directional changes in MLEs and their corresponding BoEs were generally consistent and the MLE/BoE relationship maintained. Switching to a lower yield cigarette generally resulted in reductions in exposure with the resultant exposure level being similar to that seen in regular smokers of the lower yield cigarette.

© 2011 Elsevier Inc. All rights reserved.

1. Introduction

The ability to estimate the exposure of smokers to cigarette smoke constituents is critical for the assessment of changes in cigarette design aimed at reducing smokers' exposure to cigarette smoke.

Biomarkers of exposure (BoE) to smoke constituents have been developed and used in a number of smoke exposure studies. These include the measurement of the levels of nicotine in venous blood (Russell et al., 1980,1986; Ebert et al., 1983; Gori and Lynch, 1985), cotinine in venous blood (Gori and Lynch, 1985; Bridges et al., 1990; Rosa et al., 1992), cotinine in saliva (Jarvis et al., 2001) and

* Corresponding author. Address: British American Tobacco, Group Research and Development, Regents Park Road, Millbrook, Southampton SO15 8TL, UK. Fax: +44 (0) 23 8079 3076.

E-mail address: jim_shepperd@bat.com (C.J. Shepperd).

0273-2300/$ - see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.yrtph.2011.05.011

multiple nicotine metabolites in urine (Byrd et al., 1998; Ueda et al., 2002). More recently studies have been conducted on the measurement of BoE to cigarette smoke toxicants such as NNK (4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone) and pyrene (Hecht et al., 2005; Benowitz et al., 2005; Scherer et al., 2007; Men-des et al., 2008, 2009; Shepperd et al., 2009), benzene (Scherer et al., 2006, 2007; Mendes et al., 2008), acrolein (Scherer et al., 2006, 2007; Mendes et al., 2008; Shepperd et al., 2009), and 1,3-butadiene (Scherer et al., 2006; Mendes et al., 2009). Most of the biomarker studies rely on the measurements of metabolites of the smoke constituent. Smokers may differ markedly in their metabolism of smoke constituents. Additionally, the commonly measured biomarkers have different elimination half-lives, some relatively short e.g., nicotine, others relatively long e.g., NNAL (4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol) (metabolite for NNK). Consequently, metabolism and half-life issues (where time since exposure would affect biomarker level) may create problems

in attempting to assess exposure to cigarette smoke from metabolites in urine and saliva. However, problems caused by inter-subject variability in metabolism can be minimised by using each subject as their own control during a cigarette brand-switch study. Elimination half-life problems can be minimised by providing sufficient durations of biomarker monitoring during the smoking of the pre- and post-switch cigarettes.

Another approach to estimating exposure to smoke constituents is based on the analysis of spent cigarette filters. This method relies on the measurement of the nicotine or solanesol content of the filter after smoking, and the relationships between nicotine or solanesol in the filter and the levels of smoke constituents in mainstream smoke when the cigarette is smoked using a wide range of machine-smoking regimes. From these two pieces of information it is possible to estimate the amount of nicotine or other smoke constituent that exited the filter and entered the smoker's mouth -Mouth Level Exposure (MLE). However, filter analysis methods may be limited in their ability to predict exposures of smoke constituents to, and absorption from, the respiratory tract as they do not take account of non-inhaled puffs, 'mouth-spill' post-puff inhalation, exhalation patterns and differences in the respiratory retention of individual smoke constituents. Filter analysis methods have previously been shown by a number of authors to provide good estimates of MLE (St. Charles et al., 2006, 2010; Polzin et al., 2009; Shepperd et al., 2006, 2009; Ashley et al., 2010). Recently, in their review of cigarette filter based assays as proxies for smoke exposure Pauly et al. (2009) concluded that filter based methods may have utility as proxy measures of mouth level exposure in clinical studies.

The findings presented in this manuscript extend those previously reported (Shepperd et al., 2009) which used data derived from the same study. Our earlier paper focused on the primary objective of the study, which was to compare smokers' MLEs and BoEs to specific cigarette smoke constituents as determined by the use of both the filter analysis method (MLEs) and the analysis of biomarkers of exposure in blood, urine and saliva (BoEs). We demonstrated that these two exposure estimates were significantly correlated and we concluded that the measurement of MLEs by filter analysis may provide a simple and effective alternative to BoEs for estimating smokers' exposure to smoke constituents. This earlier paper (Shepperd et al., 2009) used data from the control groups of smokers who smoked a brand of cigarette throughout the study with the same or similar ISO tar yield as their usual cigarette (10,4 or 1 mg ISO tar yield). However, this current report addresses the data from two additional groups of smokers who were switched from their usual yield product (10 or 4 mg ISO tar) to a lower ISO tar yield cigarette (4 or 1 mg ISO tar respectively) during the study. This switching aspect was incorporated into the study design in order to improve our understanding of the relationship between MLEs and BoEs. The hypothesis underlying this approach was that the switch from one ISO tar yield to a lower ISO tar yield would change smoke exposure. Upsetting the equilibrium or steady state of smoke constituent exposure using an ISO tar yield switch allows the direction and magnitude of the measured change in exposure estimates from the two methods to be compared; if there is a change in exposure as a result of switching, is the direction and magnitude of the change equally demonstrated by both methods?

The data presented here, which includes the previously published findings, allows both an examination of the influence of ISO tar yield on estimated exposure (Shepperd et al., 2009) and the effect of a switch to lower ISO yield on estimated exposure. Whilst this was not the primary objective of the current study, the inclusion of the control groups means that the study design permits comparisons to be made that contribute to this discussion on the relationship between ISO tar yield and exposure to cigarette

smoke constituents. In the previously reported aspects of this study (Shepperd et al., 2009), smokers of lower yield cigarettes generally had lower levels of MLE and biomarkers than those smoking higher yield cigarettes. This second paper not only describes any changes in exposure that might occur following a short-term switch to a lower ISO yield cigarette but it is also possible to compare levels of exposure seen in those smokers who did not switch, or those who usually smoke the lower ISO yield cigarette.

There have been a number of published studies on the effect of switching from higher to lower ISO/FTC yield cigarettes on exposure to nicotine (see Scherer, 1999). More recently there have been a few brand-switching studies on the exposure to nicotine and smoke toxicants such as NNK and pyrene (Benowitz et al., 2005, 2009; Mendes et al., 2008). Benowitz et al. (2005) reported small reductions in blood cotinine and carboxyhaemoglobin levels expressed on a per cigarette basis but no changes in the levels of biomarkers for NNK and pyrene when smokers switched from their regular cigarette to one 50% lower in FTC nicotine yield for a period of 1 week. A reduction in the levels of urinary nicotine metabolites was also reported by Mendes et al. (2008) when smokers switched from their regular to lower yield cigarettes in the short (8 days) and long-term (24 weeks) phases of their study. However, in contrast to the results of Benowitz et al. (2005), Mendes et al. also reported short and long-term reductions in the urinary biomarkers for NNK and pyrene when smokers switched from a 15 to a 6 mg FTC tar yield cigarette. In their more recent paper, Benowitz et al. (2009) reported significant reductions in the levels of cotinine in plasma and NNAL and 2-naphthol in urine when smokers switched from their usual brand (average yield of 1.05 mg nicotine) to a 0.1 mg nicotine yield brand. However, no significant changes were seen in these biomarkers when smokers switched from their usual brand to a 0.6 mg nicotine yield brand. The switching aspect of our study provides a source of additional information on the effect of brand-switching on smokers exposure to smoke constituents such as NNK, pyrene and acrolein.

2. Materials and methods

2.1. Study design

Three hundred healthy volunteers were recruited into the study: 250 smokers and 50 non-smokers. The 250 smokers were allocated to one of five groups according to the ISO tar yield of their usual cigarette brand. Groups 1 and 2 each contained 50 smokers of 9-10 mg ISO tar yield cigarettes; groups 3 and 4 each contained 50 smokers of 4-5 mg ISO tar yield cigarettes and group 5 contained 50 smokers of 1-2 mg ISO tar yield cigarettes. The 50 non smokers were assigned to group 6.

Three commercial brands of cigarettes were used for this study, yielding nominally 10, 4 and 1 mg ISO tar thereby spanning the full range of ISO tar yields legally available in Germany at the time of the study. Smokers were supplied with cigarettes of the same tobacco blend style as their normal brand for the duration of the study and all smokers began the study on a similar ISO tar and nicotine yield product to their usual brand. Although smokers were required to change to a single brand at the start of the study, in all cases the change was to a cigarette that was very similar in all respects to their usual brand, and subjects were given 5 days acclimatisation on the supplied cigarettes before baseline measures were taken. Any behavioural changes that may have remained would not have influenced the objectives of this study which were to determine whether the filter analysis and biomarker methods would detect the same direction and magnitude of any resultant change in exposure following a switch to a lower yield cigarette.

Subject recruitment and all clinical aspects of this study were conducted by MDS Pharma Services, Hamburg, Germany. To ensure compliance, all study samples (biological samples and spent filters) were collected during the clinical periods only. Analyses of urinary biomarkers were conducted by MDS Pharma Services laboratories in Sittingbourne, UK and Zurich, Switzerland. Analyses of plasma and saliva biomarkers were conducted by Analytisch Biologisches Forschungslabor GmbH, Munich, Germany. Cigarette filters were analysed by British American Tobacco, Group Research & Development (BAT GR&D), Southampton, UK.

The study protocol and informed consent forms were approved by the Ethics Committee of the Ärztekammer Hamburg and the clinical study was conducted in accordance with the World Medical Association Declaration of Helsinki (World Medical Association 2004) and International Conference on Harmonisation (ICH) Guidelines for Good Clinical Practice (GCP) (International Conference on Harmonisation 1996), current as at the date of the study. Based on a recommendation by MDS Pharma Services, which was endorsed by the ethics committee, it was agreed that subjects would be paid a stipend of 1500 Euros for their participation in the study.

2.1.1. Control groups

Three of the five smoker subject groups were control groups who smoked a single brand for the entire study. Group 1 smoked the 10 mg ISO tar yield cigarette, Group 3 smoked the 4 mg ISO tar yield cigarette and Group 5 smoked the 1 mg ISO tar yield cigarette. The results from these three groups have been reported and discussed previously (Shepperd et al., 2009). Some data from these groups are included here where they serve as non-switching control groups and therefore provide a better understanding of any changes that might occur in the two groups switching from higher to lower ISO tar yield cigarettes.

2.1.2. Switching groups

The two remaining smoker groups began the study smoking the supplied brand that was similar to their usual ISO tar and nicotine yield. They were then switched to a lower ISO tar and ISO nicotine yield cigarette for a period of 12 days. Group 2 started the study smoking the 10 mg ISO tar yield cigarette before switching to the 4 mg ISO tar yield cigarette. Group 4 started the study smoking the 4 mg ISO tar yield cigarette before switching to the 1 mg ISO tar yield cigarette. These two groups of smokers are the main focus of this paper.

The study design is summarised in Fig. 1. It includes both ambulatory and clinical confinement periods. The ambulatory periods allowed for acclimatisation to the product prior to clinical confinement and sample collection. The clinical periods were required to facilitate complete and accurate collection of 24 h urine samples and filter tips from the smoked cigarettes. It also ensured that subjects were compliant with the protocol in terms of diet and cigarette consumption.

Full details of subject screening, ambulatory visit events and clinical confinement events are detailed by Shepperd et al. (2009).

The study design ensured that there were no mixed groups and at any one time all subjects in the clinic who were smokers were supplied with the same ISO tar yield cigarette. Potential confounding effects were minimised or controlled as previously reported (Shepperd et al., 2009) with the non smoker group providing measures of biomarkers that arise from non tobacco sources.

2.2. Product selection

Three commercial cigarette products were selected as previously described (Shepperd et al., 2009). The blend styles and key physical parameters are shown in Table 1 together with the

mainstream smoke yields measured under both ISO (ISO 2000) and Massachusetts Intense (Commonwealth of Massachusetts, 1997) regimes.

2.3. Subject selection

All subjects enrolled in this study were judged by the Principal Investigator (PI) to be normal, healthy volunteers who met all inclusion and none of the exclusion criteria as previously described (Shepperd et al., 2009).

2.4. Methods

The methods used were as previously reported in detail by Shepperd et al. (2009) including details on: - home smoking; filter collection and analysis (including calibration for nicotine, NNK, pyrene and acrolein MLE); urine collection and biomarker analyses; saliva and plasma collection and analyses.

2.4.1. Home smoking

Smokers were supplied with cigarettes for home smoking during the ambulatory sections of the study. The number of cigarettes given to each subject was based on their claimed daily consumption. Subjects were asked to record actual daily consumption during ambulatory periods in a diary.

2.4.2. Filter collection and analysis

Smokers were permitted to smoke ad libitum throughout the study which, in the clinic, was confined to a designated room with air filtration systems installed. Smokers were issued with a cigarette on request, with the next cigarette only supplied on receipt of the spent cigarette filter, thus ensuring complete collection of all filters in a 24 h period. On receipt, clinic staff cut a 10 mm section from the mouth-end of the filter using a custom filter cutter. This section was retained in an airtight aluminium tin which was stored for up to 24 h at room temperature before being dispatched by air freight to BAT for analysis. The times during which filters were collected for analysis are indicated in Fig. 1.

MLEs to nicotine, NNK, pyrene and acrolein were estimated using methods fully described by Shepperd et al. (2009). Briefly, the three cigarette types were calibrated by establishing the relationships between the amounts of nicotine retained in the filters and the mainstream smoke yields of the relevant constituents during machine smoking under a range of smoking regimes. The amounts of nicotine retained in the filters of the cigarettes smoked by the subjects were measured and MLEs for the four constituents were determined using the calibration relationships for each constituent and each cigarette type.

2.4.3. Urine collection and analysis

For all subjects 24 h urine samples were collected on the days indicated in Fig. 1. The samples were analysed for total nicotine, total cotinine, and total trans-3-hydroxycotinine; total 4-(methylnit-rosamino)-1-butanol (NNAL); hydroxypropyl mercapturic acid (HPMA) and total 1-hydroxypyrene (1-OHP). 'Total' refers to the sum of the free and conjugated forms of the biomarker. Full details of the urine collection and analysis methods are shown in Shep-perd et al. (2009).

2.4.4. Saliva and plasma collection and analyses

Saliva and blood samples were obtained from the subjects at approximately 07:00 and 17:00 h on the days indicated in Fig. 1. Full details of the sampling methods are contained in Shepperd et al. (2009). The saliva and plasma samples were analysed for coti-nine using the method described by Scherer et al. (2007).

Figure 1 Study Design

■ Study day

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Switch

Switch

Key: -

24hrs at home - diary kept 24hrs in clinic ^^J 24hrs in clinic - filter collection (smokers only) toJ 24hrs in clinic - filter collection (smokers only), biofluid collection (all)

Fig. 1. Study design.

Table 1

Cigarette product details.

Group 1 and 2 2, 3 and 4 4 and 5

Subject chosen product

ISO tar, mg/cig 9-10 4-6 1-2

Blend style US US US

Format King size* King size* King size*

Supplied commercial product

Blend style US US US

Format King size* King size* King size*

Pack tar/nicotine/CO yields, mg/cig 10/0.8/10 4/0.4/5 1/0.1/2

ISO yields

Tar, mg/cig 10.5 4.7 1.0

Nicotine, mg/cig 0.83 0.45 0.12

CO, mg/cig 10.8 5.8 1.7

NNK, ng/cig 28.8 11.5 <LOQ*

Acrolein, ig/cig 45.3 27.3 7.3

Pyrene, ng/cig 49.4 33.0 11.2

Massachusetts yields

Tar, mg/cig 24.0 13.6 7.0

Nicotine, mg/cig 1.80 1.17 0.71

CO, mg/cig 22.3 15.5 11.2

NNK, ng/cig 65.6 33.6 19.7

Acrolein, ig/cig 108.0 77.2 53.1

Pyrene, ng/cig 105.3 72.2 43.0

Design

Filter ventilation, % 24 42 75

Paper permeability, CU 51 57 51

Total open pressure drop mmWG 100 101 86

CU - coresta units. WG - water guage.

# King size approximate dimensions; length 83 mm, circumference 24.75 mm, filter length 27 mm.

* <LOQ - less than the limit of quantitation for NNK (8 ng/cig).

2.4.5. Statistical analyses

Minitab® (version 15, Minitab Ltd., UK) was used for statistical analyses. For each subject a mean of the results for each period were used (i.e. a mean of day 6 and day 7 results for period 1 and a mean of day 18 and day 19 results for period 2) in the

analyses for both MLEs and BoEs . The cigarettes per day, MLE and BoEs results expressed both per cigarette and per day were summarised by group and period using Minitab's descriptive statistics. Checks for normality were carried out, using normality plots and standardised residuals, before comparison between periods for each group was made using paired t-tests with Bonferroni correction.

Modified, combined 'box plots' were constructed to show the relationship between control group results and switching group results. They combined the first and third quartile results plus the mean for each relevant MLE/biomarker pair in a two dimensional plot with MLE on the x-axis and biomarker on the y-axis. Separate boxes represent each group/period demonstrating the relationship between the relevant biomarker and constituent MLE. The use of boxes gives a representation of the spread of data within each group/period.

The mean daily cigarette consumption rates, MLEs and BoEs for period 1 and 2 for the switching groups were compared with those for their corresponding non-switching, control groups using a two sample t-test with unequal variances.

The relationship between MLE, as determined by filter analysis, and the relevant BoE was assessed by linear regression.

3. Results

A total of 300 subjects were recruited for this study. Results from 280 subjects (47 in group 1, 46 in group 2, 45 in group 3, 44 in group 4, 48 in group 5 and 50 in group 6) were used in the final data analysis. Data from subjects not used were for 6 subjects who chose to withdraw from the study, 3 subjects who were removed for violating protocol, 2 subjects who were withdrawn from the study due to serious adverse events and data from 9 subjects who completed the study but were excluded due to errors in sample collection.

Subject demographics and self-reported smoking habits are shown in Table 2.

The mean and 95% confidence interval (CI) results obtained for the MLEs per day and BoEs, along with cigarette consumption are summarised in Table 3 by group and by period. A paired t-test with Bonferroni correction was used to compare results between

Table 2

Demographics and self-reported smoking habit.

Group No. subjects Gender Age Cigs per day Pack nicotine (mg/cig) Pack tar (mg/cig)

n (% male) Mean (min-max) Mean (min-max) Median (min-max) Median (min-max)

1 47 68 42(22-72) 21 (14-30) 0.8 (0.7-0.9) 10.0 (10.0-10.0)

2 46 63 33 (21-60) 21 (14-30) 0.8 (0.7-0.9) 10.0 (9.0-10.0)

3 45 44 36 (21-66) 20 (14-30) 0.5 (0.4-0.6) 6.0 (4.0-6.0)

4 44 34 35 (21-59) 20 (6-30) 0.6 (0.4-0.6) 6.0 (4.0-6.0)

5 49 37 37 (22-70) 18 (8-30) 0.1 (0.1-0.2) 1.0 (1.0-2.0)

6 50 62 44 (22-74) 0 n/a n/a

Means shown for age and cigs per day (cigarettes per day) since continuous variables. Medians shown for pack nicotine and pack tar since categorical variables.

periods for each MLE or BoE by group and the p-values are included in this table. Since there were 12 comparisons (i.e. nicotine, NNK, acrolein and pyrene MLE, plus the urinary biomarkers, plus 2 plasma and 2 saliva cotinine measures) then the Bonferonni correction for the p value for significance was 0.05/12 = 0.004 and therefore comparisons with p values > 0.004 were not significant.

Mean daily cigarette consumption was higher in period 2 than in period 1 for all five smoker groups. Closer examination of cigarette consumption in the clinic identified a systematic increase

across the study over time, particularly for the last clinic day (day 19) where 87% of subjects showed increased cigarette consumption compared with baseline (day 6). Fig. 2 shows the mean cigarette consumption per day for each group and demonstrates this systematic observation of increasing cigarette consumption across the study for all groups.

It is unlikely that this trend in increased consumption as the study progressed occurred as a consequence of the switch from higher to lower yield cigarettes because increased consumption

Table 3

Mean (95% CI) Mouth Level Exposure (MLE), biomarkers of exposure (BoE) and cigarette consumption by group expressed as amount per day.

Group N Period 1 Period 2 p-Value Period 1 Period 2 p-Value

Cigarettes per day

1 47 21.1 (19.6, 22.6) 23.8 (22.3, 25.2) 0.000

2 46 21.8 (20.4, 23.1) 25.3 (23.1, 27.5) 0.000

3 45 19.8 (18.1, 21.5) 23.2 (21.3, 25.1) 0.000

4 44 18.5 (17.2, 19.8) 22.3 (20.6, 24.0) 0.000

5 48 18.5 (16.4, 20.6) 20.1 (18.1, 22.1) 0.001

Nicotine MLE, mg/day TNeq, mg/day

1 47 30.8 (27.5, 34.1) 34.4(31.1,37.7) 0.000 18.0(15.8, 20.1) 18.3 (16.4, 20.2) 0.412

2 46 29.1 (26.3, 31.9) 23.9 (21.3, 26.4) 0.000 16.9 (14.3, 19.4) 15.8 (13.7, 17.9) 0.077

3 45 19.1 (17.0, 21.3) 23.6 (20.9, 26.4) 0.000 12.4 (10.7, 14.1) 14.4 (12.2, 16.5) 0.000

4 44 18.4 (16.5, 20.3) 19.9 (17.2, 22.6) 0.032 11.8 (10.2, 13.4) 9.5 (8.1, 10.8) 0.000

5 48 15.2 (12.7, 17.8) 16.7 (14.3, 19.1) 0.002 7.1 (5.9, 8.3) 8.3 (7.0, 9.6) 0.001

NNK MLE, ng/day NNAL, ng/day

1 47 1112 (989, 1235) 1241 (1119, 1363) 0.000 491 (426, 555) 487 (424, 549) 0.682

2 46 1045 (942, 1149) 673 (599, 747) 0.000 437 (348, 526) 337 (282, 391) 0.000

3 45 540 (476, 603) 671 (590, 752) 0.000 282 (238, 325) 308 (254, 363) 0.006

4 44 522 (467, 578) 529 (464, 594) 0.693 266 (228, 304) 259 (209, 309) 0.569

5 48 411 (348, 473) 450 (390, 509) 0.001 179 (147, 211) 212 (178,245) 0.000

Acrolein MLE, ig/day 3-HPMA, jg/day

1 47 1815 (1610, 2020) 2025 (1822, 2228) 0.000 2007 (1732, 2282) 2050(1780, 2320) 0.398

2 46 1699 (1526, 1871) 1553 (1383, 1722) 0.001 1789 (1502, 2076) 1991 (1654, 2329) 0.002

3 45 1245 (1100, 1389) 1544 (1360, 1729) 0.000 1256 (1060, 1452) 1451 (1194, 1708) 0.001

4 44 1204 (1076, 1331) 1503 (1294, 1712) 0.000 1367 (1145, 1590) 1119 (916, 1323) 0.000

5 48 1147 (954, 1340) 1259 (1076, 1442) 0.002 880 (717, 1043) 988 (816, 1160) 0.007

Pyrene MLE, ng/day 1-OHP, ng/day

1 47 1809 (1615, 2002) 2020(1828, 2211) 0.000 335 (286, 384) 326 (285, 367) 0.327

2 46 1711 (1548, 1874) 1512 (1358, 1666) 0.000 303 (256, 351) 292 (250, 334) 0.314

3 45 1207 (1078, 1337) 1482 (1318, 1645) 0.000 254 (223, 284) 271 (233, 309) 0.072

4 44 1160 (1047,1272) 1180(1027, 1333) 0.597 231 (201, 261) 185 (162, 209) 0.000

5 48 910 (765, 1054) 997 (860, 1135) 0.002 147 (129, 165) 164 (145, 184) 0.017

Plasma cotinine 0700 h, ng/ml Plasma cotinine 1700 h, ng/ml

1 47 262 (233, 291) 275 (250, 301) 0.043 269 (241, 297) 290 (265, 315) 0.000

2 46 244 (212, 276) 233 (205, 261) 0.094 260 (229, 290) 247 (219, 276) 0.044

3 45 169 (143, 194) 192 (166, 218) 0.000 181 (156, 207) 204(178, 229) 0.000

4 44 172 (151, 193) 159 (138, 180) 0.029 188 (166, 210) 168 (145, 191) 0.001

5 48 130 (109, 151) 138 (115, 160) 0.088 134 (113,155) 142 (119, 164) 0.033

Saliva cotinine 0700 h, ng/ml Saliva cotinine 1700 h, ng/ml

1 47 323 (280, 366) 327 (287, 368) 0.881 342 (293, 391) 371 (335, 407) 0.099

2 46 273 (228, 319) 283 (242, 325) 0.748 338 (292, 384) 334 (289, 379) 0.711

3 45 168 (140, 195) 206 (176, 236) 0.073 215 (184, 246) 264 (226, 302) 0.000

4 43a 167 (144, 190) 180 (152, 208) 0.470 224 (194, 254) 222 (189, 254) 0.964

5 48b 136 (108, 164) 157 (128, 185) 0.315 160 (134, 187) 180(150,210) 0.001

a Saliva 1700 h pi N = 44. b Saliva 0700 h pi N = 47.

Daily Cigarette Consumption

Table 5

A comparison of mean daily exposure data for group 4 (4-1 mg switchers) relative to group 3 (4 mg control group) before (period 1) and after switching (period 2).

Fig. 2. Daily cigarette consumption by group and day.

was also observed in the non-switching groups. It is possible that this increase was due to the subjects becoming more at ease with the experimental procedures as the study progressed. Although not reported, a similar level of increase in cigarette consumption on the last day was seen in a previous clinical study with a similar experimental design (St. Charles et al., 2006; personal communication).

One data analysis approach to minimising the possible effect of the experimental conditions causing an increase in daily cigarette consumption from the start to the end of the study, thereby influencing the mean exposure per day data, is to express mean exposure per day data for the switching groups relative to the appropriate control groups for periods 1 and 2 of the study. The daily exposures for group 2 relative to group 1 are show in Table 4, and those for group 4 relative to group 3 are shown in Table 5.

During period 1 both the control (group 1) and switching group (group 2) smoked the 10 mg tar yield product. The mean daily exposure levels were marginally lower for group 2 than for group 1 during this period. However, none of these differences were statistically significant (p > 0.05 for all comparisons). Switching from the 10 to the 4 mg product resulted in mean values for group 2 being considerably lower than those for group 1 for most of the exposure parameters. These differences were statistically significant for nicotine, NNK, acrolein and pyrene MLEs (p < 0.001 in all

Table 4

A comparison of mean daily exposure data for group 2 (10-4 mg switchers) relative to group 1 (10 mg control group) before (period 1) and after switching (period 2).

Period 1 (%) Period 2 (%)

Cig consumption +3.0 +6.2

Nicotine MLE -5.6 -30.6***

TNeq -6.1 -13.5

Plasma cotinine 0700 -6.8 -15.3*

Plasma cotinine 1700 -3.5 -14.8*

Saliva cotinine 0700 -15.3 -13.4

Saliva cotinine 1700 -1.2 -9.9

NNK MLE -6.0 -45.8***

NNAL -11.0 -30.8***

Acrolein MLE -6.4 -23.3***

3-HPMA -10.8 -2.9

Pyrene MLE -5.4 -25.1***

1-OHP -9.5 -10.5

Period 1 (%) Period 2 (%)

Cig consumption -6.6 -3.8

Nicotine MLE -3.5 -15.7

TNeq -4.8 -34.1***

Plasma cotinine 0700 +2.2 -17.0

Plasma cotinine 1700 +3.7 -17.5*

Saliva cotinine 0700 -0.7 -12.5

Saliva cotinine 1700 +4.2 -16.2

NNK MLE -3.2 -21.2**

NNAL -5.6 -16.0

Acrolein MLE -3.3 -2.7

3-HPMA +8.8 -22.9*

Pyrene MLE -3.9 -20.4**

1-OHP -8.9 -31.6***

Note: Group 2 mean values expressed as a % of group 1 mean values. * p < 0.05 for comparison of group 1 and 2 mean values - two sample t-test (unequal variances).

*** p < 0.001 for comparison of group 1 and 2 mean values - two sample t-test (unequal variances).

Note: Group 4 mean values expressed as a % of group 3 mean values. * p < 0.05 for comparison of group 3 and 4 mean values - two sample t-test (unequal variances).

** p < 0.01 for comparison of group 3 and 4 mean values - two sample t-test (unequal variances).

*** p < 0.001 for comparison of group 3 and 4 mean values - two sample t-test (unequal variances).

cases), total NNAL (p < 0.001) and both measures of plasma cotinine (p < 0.05 for both).

A similar pattern to the one described above was observed when the group 4 (switchers from 4 to 1 mg product) and group

3 (4 mg control group) mean daily exposure data were compared (Table 5). There were marginal and statistically insignificant differences (p > 0.05 for all comparisons) in the MLEs and BoEs between the two groups during period 1 when both groups smoked the

4 mg product. However, in period 2 the mean MLEs and BoEs for the switchers were consistently lower that of the control group except for acrolein MLE. These differences were statistically significant for NNK and pyrene MLEs (p < 0.01 for both), TNeq and 1-OHP (p < 0.001 for both), 3-HPMA and plasma cotinine 1700 h (p < 0.05 for both). The difference in nicotine MLE just failed to reach significance at the 95% level.

The exposure data can also be expressed on a per cigarette basis. The MLE data were initially measured on a per cigarette basis and were converted to MLE per day by multiplying with daily cigarette consumption. However, the urinary biomarker data reflect daily exposure. Hence exposure per cigarette values were obtained by simply dividing the BoE values by the corresponding cigarette consumption value. The exposure per cigarette data are summarised in Table 6 and paired t-tests with Bonferroni correction used to provide p values. Again, as with Table 3, p < 0.004 are required to indicate a significant difference.

Comparing MLE estimates for periods 1 and 2, there were no significant differences (p > 0.004) for the control groups 1 and 5 while control group 3 showed barely significant increases (p = 0.003) for nicotine, NNK, acrolein and pyrene. Significant reductions in levels of BoEs were found for the 10 mg control group (group 1) for TNeq, NNAL, 3-HPMA and 1-OHP but no change in plasma or saliva cotinine. There were no significant differences in any BoEs for either of the other two control groups.

In contrast, switching from the 10 to the 4 mg cigarette (i.e. group 2) produced significant reductions in all MLEs per cigarette for nicotine, NNK, acrolein and pyrene. There were also significant reductions in the levels of BoEs per cigarette for TNeq, NNAL, 1-OHP, plasma cotinine, and saliva cotinine 1700 h. There were no significant changes in the levels of 3-HPMA or saliva cotinine 0700 h.

Switching from the 4 to the 1 mg cigarette (i.e. group 4) produced statistically significant reductions in all the measured MLEs per cigarette except acrolein MLE. Significant reductions following

Table 6

Mean (95% CI) Mouth Level Exposure (MLE) and Biomarkers of exposure (BoE) by group and period expressed as amount per cigarette.

Group N Period 1 Period 2 p-Value Period 1 Period 2 p-Value

Nicotine MLE, mg/cig TNeq, mg/cig

1 47 1.44 (1.35, 1.54) 1.44 (1.35, 1.52) 0.537 0.84 (0.76, 0.92) 0.77 (0.71, 0.84) 0.002

2 46 1.33 (1.24,1.41) 0.94 (0.89, 1.00) 0.000 0.76 (0.66, 0.85) 0.62 (0.56, 0.68) 0.000

3 45 0.95 (0.88, 1.03) 1.00 (0.92, 1.08) 0.003 0.62 (0.54, 0.69) 0.60 (0.54, 0.66) 0.301

4 44 0.98 (0.92, 1.04) 0.88 (0.80, 0.95) 0.000 0.62 (0.56, 0.69) 0.42 (0.37, 0.46) 0.000

5 48 0.80 (0.73, 0.86) 0.81 (0.74, 0.87) 0.509 0.38 (0.33, 0.43) 0.41 (0.36, 0.45) 0.272

NNKMLE, ng/cig NNAL, ng/cig

1 47 52.1 (48.6, 55.6) 51.8 (48.5, 55.0) 0.537 22.9 (20.6, 25.2) 20.3 (18.2, 22.3) 0.000

2 46 47.7 (44.5, 50.8) 26.6 (24.8, 28.3) 0.000 19.8 (16.0, 23.6) 13.3 (11.6,15.0) 0.000

3 45 26.9 (24.5, 29.2) 28.4 (25.9, 30.8) 0.003 14.0(12.1, 16.0) 12.8 (11.1, 14.5) 0.007

4 44 27.7 (25.9, 29.5) 23.3 (21.6, 25.0) 0.000 14.0 (12.4, 15.6) 11.3 (9.7, 12.9) 0.000

5 48 21.6 (20.1, 23.1) 21.8 (20.4, 23.3) 0.509 9.7 (8.1, 11.3) 10.5 (9.2, 11.8) 0.158

Acrolein MLE, ig/cig 3-HPMA, ig/cig

1 47 85.0 (79.0, 91.0) 84.5 (78.9, 90.0) 0.537 92.6 (83.3, 101.9) 85.4 (76.3, 94.5) 0.003

2 46 77.4 (72.0, 82.8) 61.3 (57.4, 65.2) 0.000 80.0 (69.3, 90.6) 76.2 (67.7, 84.8) 0.047

3 45 62.0 (56.7, 67.2) 65.3 (59.8, 70.9) 0.003 62.5 (54.3, 70.6) 59.3 (51.8, 66.8) 0.090

4 44 63.9 (59.7, 68.0) 66.0 (60.1, 71.9) 0.181 71.4 (62.5, 80.3) 48.6 (42.2, 55.0) 0.000

5 48 59.9 (54.7, 65.2) 60.7 (55.5, 65.8) 0.509 46.4 (40.6, 52.1) 47.4 (41.6, 53.1) 0.630

Pyrene MLE, ng/cig 1-OHP, ng/cig

1 47 84.8 (79.4, 90.2) 84.3 (79.4, 89.3) 0.537 15.6(13.8, 17.3) 13.8 (12.3, 15.3) 0.000

2 46 78.0 (73.2, 82.8) 59.8 (56.7, 62.8) 0.000 13.7 (11.8, 15.6) 11.6 (10.3, 12.9) 0.000

3 45 60.3 (56.1, 64.4) 62.9 (58.5, 67.3) 0.003 13.3 (11.5,15.0) 11.8 (10.6, 13.0) 0.003

4 44 61.7 (58.5, 65.0) 51.9 (47.8, 56.0) 0.000 12.5 (11.2, 13.8) 8.4 (7.5, 9.2) 0.000

5 48 47.7 (44.1, 51.4) 48.3 (44.7, 51.8) 0.509 8.5 (7.5, 9.5) 8.6 (7.6, 9.6) 0.877

Plasma cotinine 0700 h, ng/ml/cig Plasma cotinine 1700 h, ng/ml/cig

1 47 12.4 (11.4, 13.5) 11.7 (10.7, 12.8) 0.032 12.8 (11.8,13.8) 12.4 (11.4, 13.4) 0.184

2 46 11.1 (9.8, 12.4) 9.3 (8.4, 10.3) 0.000 11.8 (10.6, 13.0) 9.9 (9.0, 10.8) 0.000

3 45 8.6 (7.3, 9.9) 8.4 (7.3, 9.5) 0.509 9.1 (7.9, 10.3) 8.9 (7.8, 10.0) 0.248

4 44 9.3 (8.3, 10.2) 7.2 (6.3, 8.0) 0.000 10.1 (9.1, 11.0) 7.5 (6.7, 8.3) 0.000

5 48 6.9 (6.1, 7.7) 6.8 (5.8, 7.8) 0.552 7.1 (6.4, 7.8) 6.9 (6.1, 7.7) 0.317

Saliva cotinine 0700 h, ng/ml/cig Saliva cotinine 1700 h, ng/ml/cig

1 47 15.2 (13.5, 16.9) 13.8 (12.3, 15.3) 0.021 16.0(14.2,17.7) 15.9 (14.4, 17.4) 0.786

2 46 12.3 (10.5, 14.1) 11.2 (9.8, 12.6) 0.009 15.3 (13.5, 17.1) 13.2 (11.7, 14.7) 0.000

3 45 8.4 (7.0, 9.7) 9.0 (7.7, 10.2) 0.194 10.8 (9.3, 12.2) 11.5 (10.0, 13.0) 0.052

4 43a 8.9 (7.8, 10.0) 8.1 (7.0, 9.1) 0.009 11.9 (10.5, 13.4) 9.9 (8.6, 11.1) 0.000

5 48b 7.0 (5.9, 8.1) 7.6 (6.4, 8.8) 0.032 8.4 (7.5, 9.3) 8.8 (7.7, 9.9) 0.221

a Saliva 1700 h pi N = 44. b Saliva 0700 h pi N = 47.

the switch to the 1 mg cigarette were also observed for all of the BoEs per cigarette except salivary cotinine 0700 h.

The effects of switching brand on both MLE and BoE results per cigarette are also shown in Fig. 3, where modified, 2-dimensional, box plots show the boundaries of the central 50% of data and the median for each group/period combining relevant MLE and urinary BoE pairs.

The nicotine MLE and nicotine biomarker data from groups 2 and 4 were used to determine the degrees of compensation that occurred following the switch from higher to lower ISO yield cigarettes. The following equation was used to determine the compensation index (Clx) (from Scherer, 1999):

Clx = 1 -[(% change in exposure marker for nicotine) /(% change in nicotine ISO yield)]

Nicotine MLE, TNeq, plasma and saliva cotinine values were used as nicotine exposure markers and % changes in lSO nicotine yields of -45.8% (group 2) and -73.3% (group 4) were used to derive the Clx values shown in Table 7.

The calculated Clx values were less than one but greater than zero in each case indicating partial compensation following the brand switches in both groups 2 and 4. The degree of partial compensation differed according to the nicotine exposure marker used in the calculations. For the biomarkers of nicotine exposure, saliva cotinine produced the highest Clx value in both groups. Nicotine MLE produced a Clx that was lower than those calculated

using the biomarkers of nicotine exposure for the group 2 smokers. However, this was not the case for the group 4 smokers as nicotine MLE produced a higher Clx than the biomarkers of nicotine exposure.

Previously (Shepperd et al., 2009) a regression analysis was used to assess the correlation between BoE and MLE. The results of a similar analysis of the per cigarette results comparing control groups to switching groups and splitting the switching groups into pre-switch (period 1) and post-switch (period 2) are shown in Table 8.

The regression analysis again shows strong positive linear relationships across all comparisons for control groups and switching groups, both pre and post-switch, with p < 0.001 for slope in all cases. The slopes and intercepts are similar comparing the same MLE/biomarker pairs with the slope and confidence interval for the control groups falling within the confidence interval for the switching results. The Pearson correlation coefficient, r, is also similar between control and switching groups for the relevant pairs. The only exception to this is the 3-HPMA/acrolein MLE pair where the intercept post-switch has decreased such that the 95% confidence intervals for the mean control group slope does not lie entirely within the interval for the post-switching results.

A key aspect of the study design is that subjects in groups 2 and 4 can be used as their own control by comparing directional changes in their MLEs with those of their corresponding BoEs following the switch from higher to lower yield cigarettes.

Table 9 summarises the directional changes in each MLE and BoE pair following the switch from higher to lower yield cigarettes in groups 2 and 4.

ig 0.8

en 0.4

f 20 CT

5 15 <

g 10 5 0

120-I 100' 80' 6040200 -

Group 2

Interquartile Range (Q1 - Q3)

10mg /

0.5 1.0 1.5

Nic MLE mg/cig

Group 2

Interquartile Range (Q1 - Q3)

20 30 40 50 NNK MLE ng/cig

Group 2

Interquartile Range (Q1 - Q3)

I__. J

40 60 80 Acrolein MLE 'g/cig

Group 2

Interquartile Range (Q1 - Q3)

Pyrene MLE ng/cig

gi 0.8

mg 0.6

en 0.4

ic/ 80-

20 20-,

18 18-

16 16-

14 ___ gi 14-

12 1 1 X g/ 12-

10 1 + 1 n 10 -

8 1___1 О. X 8-

6 О 6-

4 4-

2 2-

0 ■ .... 0-

Control Groups

Interquartile Range (Q1 - Q3)

10mg /

2. J- J

0.5 1.0 1.5

Nic MLE mg/cig

Control Groups

Interquartile Range (Q1 - Q3)

r-t IX, I

I_____

10 20 30 40 50 NNK MLE ng/cig

Control Groups

Interquartile Range (Q1 - Q3)

¡3 0.8

t 0.6 CT

0.2 0.0

gc/i 20 g/

ic/ 80-

20 40 60 80 100 120 Acrolein MLE 'g/cig

Control Groups

Interquartile Range (Q1 - Q3)

20 18 16 icg 14

cg/ 12 £ 10 X 8 ° 6 4 2 0

20 40 60 80 100

Pyrene MLE ng/cig

Group 4

Interquartile Range (Q1 - Q3)

I____I

0.5 1.0 1.5

Nic MLE mg/cig

Group 4

Interquartile Range (Q1 - Q3)

20 30 40 50 NNK MLE ng/cig

Group 4

Interquartile Range (Q1 - Q3)

40 60 80 Acrolein MLE 'g/cig

Group 4

Interquartile Range (Q1 - Q3)

Pyrene MLE ng/cig

Fig. 3. Modified boxplots showing urinary biomarker versus Mouth Level Exposure (MLE). Key: 10 mg ISO tar cigarettes, 4 mg ISO tar cigarettes, 1 mg ISO tar cigarettes. Solid line period 1 data, dotted line period 2 data. Left and right box sides - 1st and 3rd quartile for MLE. Bottom and top box sides - 1st and 3rd quartile for biomarker: x - median value for period 1 data. + median value for period 2 data. Section A shows group 2 data, B shows control groups 1, 3 and 5 data, and C shows group 4 data.

Table 7

Compensation indices following the switch from the 10 to the 4 mg cigarette (group 2) and 4 to 1 mg cigarette (group 4).

Parameter Group 2 Group 4

Nicotine MLE 0.38 0.84

TNeq 0.74 0.56

Plasma cotinine 0700 0.66 0.69

Plasma cotinine 1700 0.66 0.64

Saliva cotinine 0700 0.90 0.82

Saliva cotinine 1700 0.79 0.74

lt can be seen from Table 9 that there are good agreements (70-94%) between the directional changes for MLEs and BoEs for nicotine, NNK and pyrene. However, this was not the case for acro-lein as only 52% (group 2) and 50% (group 4) of the smokers showed consistent directional changes in both indices of acrolein exposure.

4. Discussion

The availability of methods for quantifying exposure to cigarette smoke constituents is a necessary requirement for the initial phase of an assessment of cigarette-based potential reduced-exposure products (PREPS) (Institute of Medicine 2001).

ln a previous paper (Shepperd et al., 2009), we compare two approaches to the estimation of cigarette smoke exposure in a cross-sectional study of three groups of smokers who differed in the ISO tar and nicotine yields of their usual cigarette brands. One approach was the measurement of MLEs using the filter analysis method, and the other was the measurement of BoE. We reported moderate to strong correlations for nicotine MLE and urinary TNeq (r = 0.83), NNK MLE and NNAL (r = 0.76), acrolein MLE and 3-HPMA (r = 0.82), and pyrene MLE and 1-OHP (r = 0.63). These data also derive from the control groups from this current study. lt should be noted that in the earlier reported analysis, mean MLE and BoE values were calculated from all days for each subject, giving rise to a reduced n and slight differences in the values shown.

Table 8

Linear regression analysis data (per cigarette results).

Comparison n Slope (95% CI) Slope SE Intercept (95% CI) Intercept SE r

Control groups

TNeq versus nicotine MLE 280 0.52 (0.46-0.58) 0.03 0.04 (-0.03 to 0.11) 0.03 0.71

NNAL versus NNK MLE 280 0.33 (0.28-0.37) 0.02 4.0 (2.2-5.7) 0.9 0.64

3-HPMA versus acrolein MLE 280 0.99 (0.87-1.12) 0.06 -3.6 (-12.9 to 5.7) 4.7 0.68

1-OHP versus pyrene MLE 280 0.10 (0.08-0.13) 0.01 5.2 (3.3-7.1) 1.0 0.40

Plasma cotinine 0700 h versus nicotine MLE 280 7.0 (6.0-8.0) 0.5 1.6 (0.4-2.8) 0.6 0.62

Plasma cotinine 1700 h versus nicotine MLE 280 7.3 (6.4-8.2) 0.5 1.7 (0.6-2.7) 0.5 0.69

Salivary cotinine 0700 h versus nicotine MLE 279# 9.1 (7.7-10.4) 0.7 0.4 (-1.1 to 2.0) 0.8 0.61

Salivary cotinine 1700 h versus nicotine MLE 280 9.4 (8.0-10.8) 0.7 1.8 (0.2-3.3) 0.8 0.63

Switching groups - pre-switch

TNeq versus nicotine MLE 90 0.58 (0.43-0.74) 0.08 0.01 (-0.17 to 0.20) 0.09 0.62

NNAL versuss NNK MLE 90 0.38 (0.24-0.53) 0.07 2.4 (-3.5 to 8.2) 2.9 0.48

NNAL versus NNK MLE (less 1 outlier) 89 0.28 (0.18-0.37) 0.05 5.7 (2.0-9.4) 1.9 0.53

3-HPMA versus acrolein MLE 90 1.17 (0.85-1.48) 0.16 -6.8 (-30.0 to 16.3) 11.7 0.61

1-OHP versus pyrene MLE 90 0.15 (0.09-0.22) 0.03 2.5 (-2.2 to 7.1) 2.3 0.43

1-OHP versus pyrene MLE (less 1 outlier) 89 0.12 (0.06-0.17) 0.03 4.6 (0.5-8.6) 2.0 0.40

Plasma cotinine 0700 h versus nicotine MLE 90 7.0 (4.5-9.4) 1.2 2.1 (-0.8 to 5.0) 1.5 0.51

Plasma cotinine 1700 h versus nicotine MLE 90 7.3 (5.2-9.5) 1.1 2.5 (-0.1 to 5.1) 1.3 0.58

Salivary cotinine 0700 h versus nicotine MLE 89# 10.1 (7.0-13.2) 1.6 -1.0 (-4.7 to 2.7) 1.9 0.57

Salivary cotinine 0700 h versus nicotine MLE (less 1 outlier) 88 8.9 (6.2-11.7) 1.4 0.1 (-3.2 to 3.4) 1.7 0.56

Salivary cotinine 1700 h versus nicotine MLE 90 11.3 (8.0-14.5) 1.6 0.6 (-3.3 to 4.5) 2.0 0.59

Switching groups - post-switch

TNeq versus nicotine MLE 90 0.54 (0.39-0.68) 0.07 0.03 (-0.11 to 0.17) 0.07 0.61

NNAL versus NNK MLE 90 0.47 (0.29-0.64) 0.09 0.7 (-3.9 to 5.2) 2.3 0.48

NNAL versus NNK MLE (less 2 outliers) 88 0.39 (0.25-0.53) 0.07 2.2 (-1.4 to 5.7) 1.8 0.50

3-HPMA versus acrolein MLE 90 0.72 (0.39-1.05) 0.17 16.9 (-4.9 to 38.6) 11.0 0.41

1-OHP versus pyrene MLE 90 0.16 (0.10-0.22) 0.03 0.9 (-2.5 to 4.3) 1.7 0.50

1-OHP versus pyrene MLE (less 1 outlier) 89 0.14 (0.08-0.19) 0.03 2.3 (-0.9 to 5.5) 1.6 0.45

Plasma cotinine 0700 h versus nicotine MLE 90 6.2 (3.4-8.9) 1.4 2.7 (0.1-5.2) 1.3 0.42

Plasma cotinine 1700 h versus nicotine MLE 90 7.2 (4.6-9.7) 1.3 2.2 (-0.2 to 4.6) 1.2 0.50

Salivary cotinine 0700 h versus nicotine MLE 89# 9.3 (5.6-13.0) 1.9 1.2 (-2.2 to 4.7) 1.7 0.46

Salivary cotinine 1700 h versus nicotine MLE 89# 11.4 (7.6-15.3) 1.9 1.2(-2.4 to 4.8) 1.8 0.53

All groups

TNeq versus nicotine MLE 460 0.54 (0.49-0.59) 0.03 0.03 (-0.02 to 0.09) 0.03 0.70

NNAL versus NNK MLE 460 0.34 (0.30-0.39) 0.02 3.6(2.1-5.1) 0.8 0.60

3-HPMA versus acrolein MLE 460 0.98 (0.87-1.1) 0.06 -0.5 (-8.7 to 7.6) 4.1 0.62

1-OHP versus pyrene MLE 459# 0.12 (0.09-0.14) 0.01 4.3 (2.9-5.8) 0.7 0.43

Plasma cotinine 0700 h versus nicotine MLE 460 7.0 (6.1-7.8) 0.4 1.8 (0.9-2.8) 0.5 0.59

Plasma cotinine 1700 h versus nicotine MLE 460 7.4 (6.6-8.2) 0.4 1.9 (1.0-2.7) 0.4 0.66

Salivary cotinine 0700 h versus nicotine MLE 457### 9.0(7.8-10.1) 0.6 0.7 (-0.6 to 2.0) 0.6 0.58

Salivary cotinine 1700 h versus nicotine MLE 459# 9.7 (8.5-10.9) 0.6 1.9 (0.6-3.2) 0.7 0.61

CI = confidence interval. SE = standard error.

Outliers identified where standardised residual is >3.6.

# 1 Sample mising. ### 3 Sample missing.

One objective of the current study was to determine whether or not the relationships between estimates of MLEs and their appropriate BoEs were maintained when smokers were switched from their usual brand to one of a lower machine-derived smoke yield.

Using a single supplied product at each tar yield, rather than allowing individuals to source and smoke their own brands empowered the switching aspect of this study. A variety of subject sourced brands within each group would have made it difficult to control and initiate the same level of potential change for each subject. Since the objectives of this study were best served if a measurable change in smoke exposure was seen following the switch, it was also important that the cigarette brands used pre-and post-switch were significantly different, in terms of ISO yield. The previously reported results from the control groups indicated that both methods could detect a significant difference in the levels of exposure from the three ISO tar yield cigarettes and that the differences were in line with the ISO tar yield of the cigarettes smoked (Shepperd et al., 2009). In addition, using only three cigarette brands also offered several technical advantages, as previously reported (Shepperd et al., 2009).

MLE provides a measure of the amount of smoke constituent leaving the cigarette and entering the mouth during the puff process. However, factors such as possible changes in the degree of mouthspill (smoke exiting the mouth prior to post-puff inhalation) and post-puff inhalation/exhalation depths and durations following the brand-switch may disrupt the relationships between MLEs and BoEs. A change in post-puff inhalation depth and duration should only influence these relationships for those constituents whose respiratory deposition and retention characteristics are modified by changes in post-puff respiratory patterns. Of the constituents measured in the current study only the respiratory retention of NNK appears to be influenced by changes in post-puff respiratory patterns (Feng et al., 2007). Studies have shown that nicotine is very highly retained within the respiratory tract (Armit-age et al., 1975, 2004; Baker and Dixon, 2006; Feng et al., 2007), and its degree of retention is not markedly influenced by changes in post-puff inhalation depth and duration (Zacny et al., 1987; Armitage et al., 2004; Baker and Dixon, 2006; Feng et al., 2007). As both acrolein (Moldoveanu et al., 2007) and pyrene (Moldov-eanu et al., 2008) are also very highly retained in the respiratory

Table 9

Percentage of subjects where direction of change in exposure agree or disagree between methods.

Comparison Group —— (%) —+/+— (%) ++ (%) Total in agreement (%)

Nicotine MLE/TNeq 2 83 17 0 83

4 73 23 5 77

Nicotine MLE/plasma cotinine 1700 2 85 15 0 85

4 73 27 0 73

Nicotine MLE/saliva cotinine 1700 2 70 30 0 70

4 65 21 14 79

NNK MLE/NNAL 2 94 7 0 94

4 64 30 7 77

Acrolein MLE/3-HPMA 2 50 48 2 52

4 48 50 2 50

Pyrene MLE/1-OHP 2 76 24 0 76

4 75 25 0 75

Key: — Denotes a decrease in both MLE and BoE; ++ denotes an increase in both MLE and BoE; —+/+— denotes a either a decrease in MLE with an increase in BoE or vice versa.

tract, the degrees of retention of these substances are unlikely to be influenced by changes in inhalation depth and duration. In contrast, a change in the degree of mouthspill could influence the MLE/BoE relationship for all smoke constituents measured in the study.

The slopes of the regressions and the correlations (r values) of the MLEs versus corresponding BoEs (Table 8) were similar for the control and switching groups. This indicates that switching to a lower yield cigarette did not degrade the relationships between MLEs and BoEs observed when the smokers were smoking their usual brand of cigarette. This observation supports the use of either the part-filter method or conventional biomarker methods for determining the effects of brand-switching on the exposure of smokers to cigarette smoke constituents.

Further evidence for the good relationship between the two approaches to the measurement of smoke constituent exposure was the fact that for most smokers the changes in MLEs and BoEs for the majority of the measured smoke constituents were in the same direction following the switch to a lower yield cigarette (see Table 9). The exception was for acrolein exposure where directional agreement between acrolein MLE and 3-HPMA was obtained in only half of the smokers. This discrepancy between acrolein MLE and the biomarker, 3-HPMA, is also shown in Table 6. Group 2 smokers reduced their mean acrolein MLE by 20% but mean 3-HPMA was reduced by only 5% following the switch from the 10 to the 4 mg ISO tar cigarette. Mean acrolein MLE was increased in group 4 smokers by 3% whereas mean 3-HPMA was reduced by 32% following the switch from the 4 to the 1 mg ISO tar cigarette.

Acrolein is semi-volatile and is found in both particulate and vapour phase, whereas the other smoke constituents measured in the study are found predominantly in the particulate phase. The analytical method used for smoke acrolein involved a liquid trap and therefore the measured acrolein yield includes both the particulate and vapour phase components (total acrolein). This smoke yield is plotted against nicotine retained on the filter tip to provide a calibration curve to enable estimation of smoke yields from tip measurements. However, it is possible that the calibration routine used for determining MLEs for vapour phase components from the part-filter method, especially for the highly ventilated 1 mg ISO tar yield cigarette, may need to be improved.

Although, in general, the relationships between MLEs and their corresponding BoEs were maintained when smokers switched from their usual brand cigarettes to lower yield cigarettes there were differences in the magnitudes of the changes in the two types of exposure measures.

In group 2 smokers, switching from the 10 to the 4 mg ISO tar cigarettes produced greater reductions in MLEs than in their

corresponding BoEs on a per cigarette basis. For example, mean nicotine MLE was reduced by 29% whereas mean TNeq was reduced by 18%. As previously mentioned, factors such as changes in the degree of mouthspill following the switch could possibly account for these differences in magnitude. The results would be consistent with smokers reducing the degree of mouthspill when switching from the 10 to the 4 mg tar yield cigarette. However, the opposite effect was apparent in the group 4 smokers. In this instance the reduction in mean MLEs tended to be less than those for the corresponding BoEs following the switch from the 4 to the 1 mg ISO tar yield cigarette. For example, in group 4 mean nicotine MLE decreased by 10% whereas TNeq decreased by 32% following the switch. This would be consistent with the smokers increasing the degree of mouthspill following the switch to the lower yield cigarette although this seems an unlikely behavioural response. An alternative explanation of the results could be that the part-filter calibration method produced an underestimate of the MLEs for the 4 mg ISO tar cigarette and an over-estimate of the MLEs for the 1 mg ISO tar cigarette. All but 0.1% of subject tip data for the 4 mg (and 10 mg) ISO tar yield cigarettes were within calibration range whereas for the 1 mg cigarette, 18.5% were above the calibration and estimated from an extrapolation. Alternatively, measures of TNeq could provide under or overestimates. Such aspects could begin to account for the discrepancies between MLEs and BoEs in both switching groups. In a previous study Shepperd et al. (2006) compared nicotine and tar yields derived from filter analysis with yields from a range of cigarettes smoked under a variety of duplicated human puffing profiles. They reported very good agreements between the tar and nicotine yields from filter analysis and smoking machines. However, the machine smoking regimes used for calibrating the filter analysis method in that study were different to those used in the current study.

Although the main aim of the product switch was to investigate the relationships between MLEs and BoEs when smokers moved from their usual product to a lower yield brand it is worthwhile examining the data in the context of the phenomenon termed smoker compensation.

Compensatory smoking can occur when a smoker switches from a higher to a lower yield cigarette. Complete compensation occurs when there is no reduction in a smoker's exposure after the switch, and zero compensation occurs when the proportional reduction in exposure is the same as the proportional reduction in machine-derived yield. Partial compensation describes the situation when the proportional reduction in exposure is less than the proportional reduction in machine-derived yield (Scherer, 1999).

The degree of compensation can be influenced by a number of potential human smoking behaviour factors. These include changes in daily cigarette consumption, and puffing topography variables

such as puff number and puff volume (Scherer, 1999). Mean daily cigarette consumption increased when both groups of smokers (group 2 and 4) were switched from the higher to the lower yield cigarettes. However, daily consumption increases were also seen in their two non-switching, control groups (groups 1 and 3). As there were no statistically significant differences in consumption rates between the switching groups and their controls in period 2 of the study it is possible that these increases occurred as a result of the experimental design rather than as a compensatory response to the reduction in machine-derived yields in groups 2 and 4. This suggests that it might be helpful to include non-switching control groups in studies on compensatory smoking.

The compensation indices shown in Table 7 were based on nicotine exposure per cigarette data and thus were not influenced by the systematic increase in daily cigarette consumption as the study progressed. The nicotine BoE results indicated partial compensation occurred following the switches from the 10 to the 4 mg (group 2), and the 4 to the 1 mg (group 4) tar yield cigarettes. These observations are consistent with the partial compensation findings based on nicotine uptake reported in the review articles on smoker compensation (Stephen et al., 1989; Scherer, 1999) and in more recent switching studies (Mendes et al., 2008; Beno-witz et al., 2005). Our results from the group 4 smokers are also in agreement with those of a recent Benowitz et al., (2009) study which showed statistically significant reductions in nicotine exposure when smokers were progressively switched from their usual brand (12 mg FTC tar yield) to 2 mg and 1 mg FTC tar yield brands. However, Benowitz et al. (2009) did not observe a significant reduction in nicotine exposure when their subjects were switched to a 4 mg FTC tar yield brand. Thus our results from the group 2 smokers are not consistent with those from the recent Benowitz et al. (2009) study.

The compensation indices based on saliva cotinine were higher than those based on plasma cotinine for both groups of switchers. We currently have no plausible explanation for these differences.

ln addition to the reductions in nicotine MLEs and BoEs, switching from a higher to a lower yield cigarette for a period of 12 days resulted in statistically significant reductions in MLEs for NNK (both switching groups), pyrene (both switching groups) and acro-lein (group 2) together with statistically significant reductions in the BoEs, NNAL (both groups), HPMA (group 4) and 1-OHP (group 4). Mendes et al. (2008) measured 24 h urinary nicotine equivalents, plasma cotinine, NNAL, 1-OHP, 3-HPMA, S-PMA (biomarker for benzene) and COHb in smokers in short (8 days) and long-term switching (24 weeks) from US Marlboro full flavour (MFF -15 mg FTC tar yield) to either Marlboro Lights (ML -11 mg FTC tar yield) or Marlboro Ultra Lights (MUL -6 mg FTC tar yield). They compared the results from the two switching groups with those from a control group who smoked MFF throughout the study. Switching from MFF to ML produced small but statistically significant reductions in 24 h urinary nicotine equivalents and 1-OHP but no reduction in the other measured biomarkers following the short-term switch. Short-term switching from MFF to MUL resulted in statistically significant reductions in 24 h urinary nicotine equivalents, plasma cotinine, NNAL and 1-OHP, and non-significant reductions in the other biomarkers. Our 12-day switching results are broadly in line with the short-term switching data reported by Mendes et al. (2008).

Benowitz et al. (2005) conducted a short-term switching study (1 week) in which US smokers were switched from their usual cigarette to one delivering around 50% less nicotine under standard machine-smoking conditions. The average machine-derived yields for own brand cigarettes were 14.4 mg FTC tar, 1.06 mg FTC nicotine for the pre-switch brands and 6.1 mg FTC tar, 0.5 mg FTC nicotine for the post-switch brands. Benowitz et al. (2005) reported small reductions in nicotine BoEs in spot urine samples

(normalised by creatinine) but no reductions in NNAL or biomark-ers of pyrene exposure. They concluded that short-term switching to lower yield brands produced no significant reduction in carcinogen exposure. In a more recent study Benowitz et al. (2009) gradually switched smokers from their usual brand (average yields 12.1 mg tar, 1.05 mg nicotine) to five brands with progressively reduced tar and nicotine yields. Each brand was smoked for a period of 1 week and BoEs for nicotine, NNK, pyrenes (spot urine samples) and CO were measured. They reported no significant reductions in the various biomarkers for three of the cigarettes (12, 8 and 4 mg tar yield) but significant reductions in the BoEs were obtained for the 2 and 1 mg tar yield cigarettes compared with the exposures obtained from the smokers' usual brands. Our MLE and BoE results following the switch from a 4 to a 1 mg lSO tar yield cigarette are in line with those reported by Benowitz et al. (2009). However, we also observed statistically significant reductions in NNK and pyr-ene MLEs, and NNAL following a switch from a 10 to a 4 mg ISO tar cigarette whereas Benowitz et al. (2009) failed to observe differences in NNAL or BoEs for pyrene when their subjects switched down to a 4 mg FTC tar yield cigarette. Spot urine samples should have higher variability than the 24 h urine samples collected in a clinic. Therefore, this may explain why Benowitz failed to detect differences that were detected by both Mendes and the current study. An additional problem with the Benowitz et al. (2009) study is that the yields of NNK and pyrene were not measured in any of the cigarettes used in the study. Thus there is no information to show the magnitude of the differences in the machine-derived yields of these constituents between the pre- and post-switch cigarettes. ln contrast, machine-derived yields of the constituents relating to the various MLEs and BoEs were measured in our study, and the three products differed markedly in the yields of NNK and pyrene.

Another interesting observation in the data from our study is that following the switch from the 10 to the 4 mg lSO tar yield cigarettes the group 2 mean values for nicotine, NNK, acrolein, and pyrene MLEs, and TNeq, NNAL and 1-OHP were very similar to those obtained by the regular smokers of 4-5 mg lSO tar yield cigarettes (i.e. group 3 smokers). A similar outcome was seen following the switch from the 4 to the 1 mg ISO tar cigarette. In this instance post-switch mean values for nicotine, NNK and pyrene MLEs, NNAL, 3-HPMA and 1-OHP were similar to those measured in the regular 1-2 mg ISO tar yield smokers. These observations are consistent with those published by Mendes et al. (2008) who reported similar levels of nicotine BoEs in smokers who switched from Marlboro full flavour to Marlboro Lights in the short-term phase of their study to those obtained from regular smokers of Marlboro Lights measured in an earlier study (Roethig et al., 2007). Mendes et al. (2008) reported that such results supported the validity of forced switching studies to evaluate exposure to different cigarette products. Our results are consistent with the work reported by Mendes et al. (2008). These data, of course, provide no information on whether the reductions in MLEs or BoEs are associated with any reductions in health risks.

ln conclusion, our study has demonstrated that MLEs for nicotine and other smoke constituents derived from the analysis of spent filters are related to BoEs during short-term brand switch studies. Although the pre- to post-switch directional changes in MLEs and their corresponding BoEs were consistent for most of the smoke constituents measured there were some differences in the magnitudes of the changes observed using the two methods of exposure assessment. Further work is required to determine whether these anomalies occurred as a result of post-puff inhalation changes e.g., degree of mouthspill, or as a result of calibration issues particularly with highly ventilated, very low tar yield cigarettes.

Conflict of Interest Statement

This work was funded by British American Tobacco (BAT), and all authors, with the exception of Dr. Mike Dixon, are full time employees of BAT. Dr. Mike Dixon's involvement was in the capacity of a paid consultant to BAT.

Acknowledgments

We thank Dr. Gerhard Scherer and Dr. Kelley St. Charles for their input into the study design and Pamela Saunders and Madeleine Ashley for their contribution to the filter analysis and Dr. Christopher Proctor for his contribution to the draft manuscript.

References

Armitage, A.K., Dollery, C.T., George, C.F., Houseman, T.H., Lewis, P.J., Turner, D.M., 1975. Absorption and metabolism of nicotine from cigarettes. Br. Med. J. 4 (5992), 313-316.

Armitage, A.K., Dixon, M., Frost, B., Mariner, D.C., Sinclair, N.M., 2004. The effect of inhalation volume and breath-hold on the retention of nicotine and solanesol in the human respiratory tract and subsequent plasma nicotine concentration during cigarette smoking. Beitr. Tabakforsch. Int. 21 (4), 240-249. Ashley, D.L., O'Connor, R.J., Bernert, J.T., et al., 2010. Effect of differing levels of tobacco specific nitrosamines in cigarette smoke on the levels of biomarkers in smokers. Cancer Epidemiol Biomarkers Prev 19 (6), 1389-1398. Baker, R.R., Dixon, M., 2006. The retention of tobacco smoke constituents in the

human respiratory tract. Inhal. Toxicol. 18 (4), 255-294. Benowitz, N.L., Jacob III, P., Bernert, J., Wilson, M., Wang, L., Allen, F., Dempsy, D., 2005. Carcinogen exposure during short-term switching from regular to light cigarettes. Cancer Epidemiol Biomarkers Prev 14, 1376-1383. Benowitz, N.L., Dains, K.M., Hall, S.M., Stewart, S., Wilson, M., Dempsey, D., Jacob III, P., 2009. Progressive commercial cigarette yield reduction: Biochemical exposure and behavioral assessment. Cancer Epidemiol Biomarkers Prev 18, 876-883.

Bridges, R.B., Combs, J.G., Humble, J.W., Turbeck, J.A., Rehms, R., Haley, N.J., 1990. Population characteristics and cigarette yield as determinants of smoke exposure. Pharmacol. Biochem. Behav. 37,17-28. Byrd, G.D., Davis, R.A., Caldwell, W.S., Robinson, J.H., deBethizy, J.D., 1998. A further study of FTC yield and nicotine absorption in smokers. Psychopharmacology 134, 291-299.

Commonwealth of Massachusetts 1997. Tobacco disclosure act, General laws of Massachusetts, Chapter 94, Section 307B, 105 CMR 660.000, Cigarette and smokeless tobacco products: Reports of added constituents and nicotine ratings. <www.mass.gov/Eeohhs2/docs/dph/regs/105cmr660.pdf>. Ebert, R.V., McNabb, M.E., McCusker, K.T., Snow, S.L., 1983. Amount of nicotine and carbon monoxide inhaled by smokers of low tar, low nicotine cigarettes. JAMA 250, 2840-2842.

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 the retention of selected smoke constituents in the respiratory tract of smokers during cigarette smoking. Inhal. Toxicol. 19,169-179. Gori, G.B., Lynch, C., 1985. Analytic cigarette yields as predictors of smoke

bioavailability. Reg. Toxicol. Pharmacol. 5, 314-326. Hecht, S.S., Murphy, S.E., Carmella, S.G., Li, S., Jensen, J., Le, C., Joseph, A.M., Hatsukami, D.K., 2005. Similar uptake of lung carcinogens by smokers of regular, light, and ultralight cigarettes. Cancer Epidemiol. Biomarkers Prev. 14, 693-698.

Jarvis, M., Boreham, P.P., Feyerabend, C., Bryant, A., 2001. Nicotine yield from machine-smoked cigarettes and nicotine intake from smokers: evidence from a representative population survey. J. Nat. Cancer Inst. 93 (2), 134-138.

Mendes, P., Kapur, S., Wang, J., Feng, S., Roethig, H., 2008. A randomized, controlled exposure study in adult smokers of full flavor Marlboro cigarettes switching to Marlboro lights or Marlboro ultra lights cigarettes. Regul. Toxicol. Pharmacol. 51 (3), 295-305.

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

Moldoveanu, S.C., St. Charles, F.K., 2007. Differences in the chemical composition of the particulate phase of inhaled and exhaled cigarette mainstream smoke. Beitr. Tabakforsch. Int. 22, 290-302.

Moldoveanu, S.C., Coleman III, W., Wilkins, J.M., 2008. Determination of polycyclic aromatic hydrocarbons in exhaled cigarette smoke. Beitr. Tabakforsch. Int. 23, 85-97.

Pauly, J.L., O'Connor, R.J., Paszkiewicz, M., Cummings, K.M., Djordjevic, M.V., Shields, P.G., 2009. Cigarette filter-based assays as proxies for toxicant exposure and smoking behavior - A literature review. Cancer Epidemiol. Biomarkers Prev. 18 (12), 3321-3333.

Polzin, G.M., Wu, W., Yan, X., McCraw, J.M., Abdul-Salaam, S., Tavakoli, A.D., Zhang, L., Ashley, D.L., Watson, C.H., 2009. Estimating smokers' mouth level xposures to select mainstream smoke constituents from discarded cigarette filter butts. Nicotine & Tob. Res. 11, 868-874.

Roethig, H.J., Zedler, B.K., Kinser, R.D., Feng, S., Nelson, B.L., Liang, Q., 2007. Short-term clinical expaoure evaluation of second-generation electrically heated cigarette smoking system. J. Clin. Pharmacol. 47, 518-530.

Rosa, M., Pacifici, R., Altieri, I., Pichini, S., Ottaviani, G., Zuccaro, P., 1992. How the steady state cotinine concentration in cigarette smokers is directly related to nicotine intake. Clin. Pharmacol. Ther. 52, 324-329.

Russell, M.A.H., Jarvis, M., Iyer, R., Feyerabend, C., 1980. Relation of nicotine yield of cigarettes to blood nicotine concentrations in smokers. Br. Med. J. 280,972-976.

Russell, M.A.H., Jarvis, M.J., Feyerabend, C., Saloojee, Y., 1986. Reduction of tar, nicotine and carbon monoxide intake in low tar smokers. J. Epidemiol. Commun. Health 40, 80-85.

Scherer, G., 1999. Smoking behavior and compensation: a review of the literature. Psychopharmacology 145, 1-20.

Scherer, G., Urban, M., Engl, J., Hagedorn, H.-W., Riedel, K., 2006. Influence of smoking charcoal filter tipped cigarettes on various biomarkers of exposure. Inhal. Toxicol. 18, 821-829.

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

Shepperd, C.J., St. Charles, F.K., Lien, M., Dixon, M., 2006. Validation of methods for determining consumer smoked cigarette yields from cigarette filter analysis. Beitr. Tabakforsch. Int. 22, 176-184.

Shepperd, C.J., Eldridge, A.C., Mariner, D.C., McEwan, M., Errington, G., Dixon, M., 2009. A study to estimate and correlate cigarette smoke exposure in smokers in Germany as determined by filter analysis and biomarkers of exposure. Regul. Toxicol. Pharmacol. 55, 97-109.

St. Charles, F.K., Krautter, G.R., Dixon, M., Mariner, D.C., 2006. A comparison of nicotine dose estimates in smokers between filter analysis, salivary cotinine, and urinary excretion of nicotine metabolites. Psychopharmacology 189 (3), 345-354.

St. Charles, F.K., Kabbani, A.A., Borgerding, M.F., 2010. Estimating tar and nicotine exposure: human smoking versus machine smoking. Regul. Toxicol. Pharmacol. 56 (1), 100-110.

Stephen, A., Frost, C., Thompson, S., Wald, N., 1989. Estimating the Extent of Compensatory Smoking. In: Wald, N., Froggatt, P. (Eds.), Nicotine, Smoking and the Low Tar Programme. Oxford Medical Publications, Oxford, pp. 100-115.

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

Zacny, J.P., Stitzer, M.L., Brown, F.J., Yingling, J.E., Griffiths, R.R., 1987. Human cigarette smoking: effects of puff and inhalation parameters on smoke exposure. J. Pharmacol. Exp. Ther. 240, 554-564.