A longitudinal study of smokers’ exposure to cigarette smoke and the effects of spontaneous product switching

Regulatory Toxicology and Pharmacology 72 (2015) 8–16

Contents lists available at ScienceDirect

Regulatory Toxicology and Pharmacology journal homepage: www.elsevier.com/locate/yrtph

A longitudinal study of smokers’ exposure to cigarette smoke and the effects of spontaneous product switching Anthony Cunningham ⇑, Johan Sommarström, Oscar M. Camacho, Ajit S. Sisodiya, Krishna Prasad Group Research and Development, British American Tobacco (Investments) Ltd., Southampton SO15 8TL, United Kingdom

a r t i c l e

i n f o

Article history: Received 4 December 2014 Available online 14 March 2015 Keywords: Mouth level exposure Biomarkers of exposure Tobacco smoke Longitudinal Compensatory smoking behaviour Cigarette consumption

a b s t r a c t A challenge in investigating the effect of public health policies on cigarette consumption and exposure arises from variation in a smoker’s exposure from cigarette to cigarette and the considerable differences between smokers. In addition, limited data are available on the effects of spontaneous product switching on a smoker’s cigarette consumption and exposure to smoke constituents. Over 1000 adult smokers of the same commercial 10 mg International Organization for Standardization (ISO) tar yield cigarette were recruited into the non-residential, longitudinal study across 10 cities in Germany. Cigarette consumption, mouth level exposure to tar and nicotine and biomarkers of exposure to nicotine and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone were measured every 6 months over a 3 and a half year period. Cigarette consumption remained stable through the study period and did not vary significantly when smokers spontaneously switched products. Mouth level exposure decreased for smokers (n = 111) who switched to cigarettes of 7 mg ISO tar yield or lower. In addition, downward trends in mouth level exposure estimates were observed for smokers who did not switch cigarettes. Data from this study illustrate some of the challenges in measuring smokers’ long-term exposure to smoke constituents in their everyday environment. Ó 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http:// creativecommons.org/licenses/by/4.0/).

1. Introduction The long-term health risks associated with cigarette smoke exposure have been extensively studied and are well established (US Department of Health and Human Services, 1982, 1983, 1984). Epidemiologic studies have shown that the risk of developing smoking-related diseases is dose related and increases with duration of smoking and consumption and that cessation generally leads to reductions in the risks of developing disease (US Department of Health and Human Services, 2004; International Agency for Research on Cancer, 2004).

Abbreviations: ADC, average daily consumption; ANOVA, analysis of variance; CiX, compensation index; ISO, International Organization for Standardization; OHcotinine, trans-30 -hydroxycotinine; LS, least squares; LSR, Lucky Strike Red; MLE, mouth level exposure; NEQ, nicotine equivalents; NNAL, 4-(methylnitrosamino)-1(3-pyridyl)-1-butanol; NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone; TobReg, World Health Organization Study Group on Tobacco Product Regulation; TP, timepoint; WHO, World Health Organization. ⇑ Corresponding author at: Group Research and Development, British American Tobacco (Investments) Ltd., Regents Park Road, Southampton SO15 8TL, United Kingdom. Fax: +44 2380 793076. E-mail address: [email protected] (A. Cunningham).

Despite the benefits of smoking cessation, there are smokers who cannot or are unwilling to stop smoking completely and this has promoted alternative tobacco harm reduction strategies such as mandating lowering of toxicants per unit of nicotine in cigarette smoke (Burns et al., 2008) and the development of potentially reduced-exposure products (Stratton and Wallace, 2001). In 2007 the World Health Organization (WHO) Study Group on Tobacco Product Regulation (TobReg) issued a technical report on the scientific basis of tobacco product regulation, which included a list of research needs in relation to ‘‘The contents and design features of tobacco products: their relationship to dependence potential and consumer appeal’’ (WHO, 2007). Included in this list are: the need to assess whether an increase or decrease in nicotine content per unit (e.g., cigarette) would be beneficial to public health; method development to assess the effects of contents and designs on toxicity, consumer appeal and the potential for dependency; investigation of the potential effect on tobacco-use patterns of efforts to control contents, appeal and dependence potential through population surveillance studies and monitoring of unintended consequences of those policies. A challenge in investigating the effect of such public health policies on cigarette consumption and exposure arises from the fact that a smoker’s exposure is known to vary from cigarette to

http://dx.doi.org/10.1016/j.yrtph.2015.03.004 0273-2300/Ó 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

cigarette (Federal Trade Commission, 1967) and that the variation between smokers is considerably greater (Hammond et al., 2005). An additional consideration is the possibility for compensatory smoking behaviour to occur in response to smoking a cigarette of higher or lower tar and nicotine yield (Benowitz, 2001). Compensation is possible through changes in puff volume, frequency or consumption. There have been a number of published studies that have reported the extent of compensation using crosssectional studies of smokers of cigarettes across a range of machine-derived yields (Byrd et al., 1998; Hecht et al., 2005; Mendes et al., 2009) or through interventional product switching studies (Benowitz et al., 2005, 2009; Mendes et al., 2008; Shepperd et al., 2011). A review of the literature concluded that compensation was on average incomplete (partial) when switching to lower yield cigarettes (Scherer, 1999; Scherer and Lee, 2014) and was generally achieved through changes in smoking intensity rather than an increase in the number of cigarettes smoked. Very few studies have measured compensatory smoking behaviour in smokers who spontaneously switched products (Lynch and Benowitz, 1987; Muhammad-Kah et al., 2011) and at multiple timepoints. This paper reports data from a longitudinal study of over 1000 German smokers of a commercial 10 mg ISO (International Organization for Standardization) tar product in their everyday environment over a period of 3 and a half years. The primary objective of the study was to assess patterns of cigarette consumption and exposure to cigarette smoke using biomarkers of tobacco smoke exposure and mouth level exposure (MLE) estimates of tar and nicotine. In addition, we studied the effects of spontaneous product switching on cigarette consumption and exposure to cigarette smoke. Secondary objectives were to assess the extent of any compensatory smoking behaviour when smokers switched to lower tar and nicotine yield cigarettes. In studying the long-term behaviours of smokers of the same product and measuring a number of endpoints we aimed to develop a suitable methodology for monitoring the smoking habits of smokers in their everyday environment over an extended period of time. This aimed to address some of the research needs raised by TobReg (WHO, 2007) whilst providing additional information on the effects of spontaneous product switching on smokers’ cigarette consumption and exposure to smoke constituents.

2. Materials and methods 2.1. Study design Over 1000 smokers who smoked a minimum of 8 cigarettes a day of the same commercial 10 mg ISO tar yield product, Lucky Strike Red (LSR), for at least 6 months were recruited into the study; a non-residential, observational, multicentre, longitudinal study across 10 cities in Germany. The study design included unrestricted periods and three clinical visits over a 12-day fieldwork period, every 6 months (referred to as a timepoint). On day one (visit 1) of the fieldwork period, subjects were requested to smoke their own purchased cigarettes for days 2–8 and to collect cigarette remains in aluminium containers provided. On day 9 (visit 2) subjects returned to the clinic with the aluminium containers and the contents counted to provide a measure of average daily consumption (ADC). Subjects were provided with a day’s worth of their own cigarette product, a filter cutter/collector device to store part-filters from smoked cigarettes (Ashley et al., 2011) and containers for 24 h urine collection. Subjects were requested to smoke only the provided cigarettes during day 11 and to collect all urine from the second void to and including the first void of day 12. At the final visit (day 12), subjects returned the 24 h urine sample, the filter cutter/collector

9

device containing part-filters and provided a saliva sample under the supervision of a study nurse. Clinical aspects of the study were conducted by an independent contract research organisation in compliance with the principles documented in the Declaration of Helsinki (World Medical Association, 1996) and the International Conference on Harmonisation Guidelines for Good Clinical Practice (International Conference on Harmonisation, 1996). The study is registered with Current Controlled Trials (ID: ISRCTN95019245). The study protocol and informed consent form were approved by the Ethics Committee of Bayerischen Landesärztekammer. Subjects were required to provide written informed consent and were provided with details of local smoking advice centres. Full details of the study design, subject screening, and study visit activities and measurements are described by Cunningham et al. (2014). 2.2. Products Subjects were provided with one day’s supply of cigarettes of their current product, at each timepoint of the study, with each product sourced from a single manufacturing batch. Subjects were provided with cigarettes for use during the period of 24 h urine and spent filter collection to minimise variation between subjects’ products and to allow sample analysis to be conducted on the same batch of cigarettes as those smoked by the study subjects. ISO tar and nicotine yields of each product supplied were determined according to ISO standards (International Organization for Standardization, 2000). 2.3. Exposure measures Estimated mouth level exposures to tar and nicotine were obtained by analysis of spent part-filters (St. Charles et al., 2009). Daily estimates of mouth level exposure were calculated using the per cigarette estimates multiplied by the number of cigarettes smoked on day 11 of the fieldwork period. Each subject’s average daily cigarette consumption was calculated from the total number of cigarette remains collected from days 2–8 of the fieldwork period. Twenty-four hour urine samples were measured for nicotine equivalents (NEQ, the molar sum of nicotine, cotinine, trans-30 -hydroxycotinine (OH-cotinine), nicotine-N-glucuronide, cotinine-Nglucuronide and trans-30 -hydroxycotinine-O-glucuronide) (Meger et al., 2002); total NNAL (urinary 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) and NNAL-glucuronides) (Scherer et al., 2007) and creatinine (Blaszkewicz and Liesenhoff-Henze, 2012). Saliva samples were analysed for cotinine and OH-cotinine according to published methodology (Scherer et al., 2007). 2.4. Data sampling and statistical analysis Statistical analysis was conducted using SAS version 9.3 (SAS Institute, Cary, NC, USA). Data were analysed separately for each endpoint, using a linear mixed model for repeated measures analysis of variance (ANOVA) with SAS PROC MIXED, to assess group changes over time for all MLE, biomarkers of exposure and consumption measurements using least squares (LS) means and 95% confidence intervals. Smoker groups for each timepoint were defined based on categorisation of each subject according to the ISO tar yield of their current product as follows: LSR (10 mg ISO tar), 8–10 mg (8–10 mg ISO tar) and 67 mg (67 mg ISO tar). The ANOVA model included the fixed term effects of smoker group (LSR, 8–10 mg or 67 mg), timepoint, gender, age category at enrolment (21–29, 30–39, 40–49, 50–64) and smoker group by timepoint interaction. The fixed term main effects and interaction term were removed from the model if not statistically significant (p > 0.05). As a multi-centre study, the linear mixed model

10

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

included the random variable, city, which was removed from the model when not statistically significant (p > 0.05). The first timepoint (TP0) was used as baseline to account for possible subject differences at the start of the study. When the main effects of timepoint or age category were found to be significant, cumulative contrast comparisons were made with adjustment for multiple comparisons. When the interaction term was found to be significant, contrasts were made between smoker groups within each timepoint, with adjustment for multiple comparisons. Descriptive statistics were calculated for demographic characteristics and baseline consumption and exposure data. Compensation index was defined as the proportional change in a subject’s exposure measurement relative to the proportional change in ISO determined yield. A compensation index per subject was calculated when switching from LSR to a product of ISO tar yield of 7 mg or lower at adjacent timepoints and then for subsequent timepoints where the subjects remains smoking a product of ISO tar yield of 7 mg or lower, according to the equation:

Table 1 Subject demographic and baseline characteristics. Baseline timepoint

  logðexposure2 Þ  logðexposure1 Þ CiX ¼ 1  logðyield2 Þ  logðyield1 Þ where CiX is the extent of compensation and exposure1 and yield1 are the levels of the exposure endpoint and ISO yield prior to the switch, respectively, and exposure2 and yield2 the levels following the switch (Frost et al., 1995). A CiX value of zero represents no compensation, whilst a value of 1 represents complete compensation. Compensation indices were calculated for subsequent timepoints where a subject remained smoking a product of 7 mg ISO tar or lower, to assess the effect of switch duration. This approach created switching events ranging from 6 to 36 months duration. Compensation indices were analysed using a linear mixed model to assess the effect of switch duration.

Number of subjects

1011

Gender, n (%) Males Females

652 (64) 359 (36)

Age, years, n (%) 21–29 30–39 40–49 50–64 Mean (SD)

418 (41) 353 (35) 173 (17) 67 (7) 33 (9)

Consumption, mean (SD) Self-reported Average daily consumptiona Day 11b

15.8 (6.0) 13.3 (6.2) 15.7 (6.6)

Exposure data, mean (SD) MLE tar (mg/cig) MLE nicotine (mg/cig) MLE tar (mg/day) MLE nicotine (mg/day) Urinary NEQ (mg/day) Urinary total NNAL (ng/day) Salivary cotinine (ng/mL) Salivary OH-cotinine (ng/mL)

17.83 (5.79) 1.50 (0.49) 282.6 (154.4) 23.98 (13.27) 13.91 (8.95) 307.72 (226.50) 266.03 (164.84) 90.70 (73.60)

n = number, SD-standard deviation, MLE = mouth-level exposure, NEQ = nicotine equivalents, NNAL = 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol, OH-cotinine = trans-30 -hydroxycotinine. a Consumption based on all cigarettes smoked over 7 days. b Consumption based on the number of spent filter-tips collected during the 24 h collection period from the cigarettes provided.

Table 2 Smoker group and brand variant composition by timepoint.

3. Results

Timepoint

3.1. Subjects and baseline characteristics A total of 1191 subjects gave their written consent to participate in the study, of which 1088 passed inclusion criteria and were enrolled in the study. A total of 1019 subjects completed the baseline assessment (TP0), of which 1011 were included in the baseline dataset, five subjects being excluded for protocol non-compliance and a further three because of errors in sample collection. A total of 546 subjects completed the final timepoint (TP6). Subjects’ demographic characteristics, and baseline daily cigarette consumption and exposure data are shown in Table 1. 3.2. Product switching Table 2 summarises the number of subjects in each smoker group per timepoint and the total number of brand variants smoked. 3.3. Repeated measures analysis Table 3 summarises the significant effects in the linear mixed model for repeated-measures analysis of mouth level exposure, biomarkers of exposure and consumption. 3.3.1. Mouth level exposure and cigarette consumption On average values for MLE to tar and nicotine per cigarette were lower for the 67 mg group compared with the LSR and 8–10 mg groups. Values for the 67 mg group were lower than the LSR and 8–10 mg groups at each timepoint, though downward trends for the LSR and 8–10 mg groups over time reduced the magnitude of

Total subjects (n) LSR group (n) 8–10 mg groupb (n) 67 mg groupc (n) Total brand variants (n) Brand variants 8–10 mg group (n) Brand variants 6 7 mg group (n)

1a

2

3

4

5

6

820 741 39 40 30 19 10

744 614 58 72 44 27 16

662 527 62 73 39 24 14

611 469 71 71 46 26 19

582 436 72 74 50 27 22

546 395 77 74 43 22 20

a

TP1 = timepoint 1, the first timepoint to assess spontaneous product switching. 8–10 mg group are those subjects who smoked a product of 8–10 mg ISO tar (not including the LSR product). c 67 mg group are those smokers who smoked a product of 67 mg ISO tar. b

the differences. For MLE to tar estimates, statistical differences were found between the 67 mg and LSR groups at timepoints 1– 3 and between 67 mg and 8–10 mg groups at timepoints 1 and 2, with a maximum difference of 3.04 mg tar/cigarette (16%) (Fig. 1a). For MLE to nicotine estimates, statistical differences were found between the 67 mg and LSR smokers at timepoints 1 and 3, with a maximum difference of 0.20 mg nicotine/cigarette (12%) (Fig. 1b). Average daily cigarette consumption and day 11 cigarette consumption did not show significant differences between smoker groups or across the timepoints of the study. Both average daily cigarette consumption and day 11 consumption for the age group 21–29 was statistically lower than that observed for all other age groups. MLE to tar and nicotine per day was lower at each timepoint for the 67 mg group compared with the LSR group, with differences of 54.1 mg tar/day (18%) and 3.57 mg nicotine/day (15%) observed at timepoint 3 (Fig. 1c and 1d). Mean values for the 67 mg group

11

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16 Table 3 Fitted estimates and statistical significance from linear mixed model for repeated measures analysis per endpoint. Endpoint MLE tar (mg/cig)

MLE nicotine (mg/cig)

MLE tar (mg/day)

MLE nicotine (mg/day)

ADC

Day 11 consumptionc

Urinary NEQ (mg/day)

Urinary total NNAL (ng/day)

Salivary cotinine (ng/mL)

Salivary OHcotinine (ng/mL)

Smoker Groupa LSR 8–10 mg 67 mg

⁄⁄⁄

⁄⁄

⁄

N.S. – – –

⁄⁄⁄

⁄⁄

290.5 284.9 262.8

N.S. – – –

⁄⁄⁄

1.50 1.51 1.41

N.S. 23.49 23.33 21.70

⁄

18.57 18.46 16.93

12.74 13.68 11.33

378.24 414.24 332.39

290.12 300.67 256.66

94.76 99.01 82.78

Gender Female Male

N.S. – –

N.S. – –

⁄⁄

⁄

⁄

22.22 23.46

13.3 13.8

N.S. – –

N.S. – –

⁄

270.0 288.9

N.S. – –

N.S. – –

Timepoint 1 2 3 4 5 6

⁄⁄⁄

⁄⁄⁄

⁄⁄⁄

⁄

⁄⁄⁄

⁄⁄⁄

287.6 289.5 284.6 285.8 269.2 259.8

23.00 23.61 22.65 22.87 21.81 23.12

N.S. – – – – – –

⁄⁄⁄

1.51 1.51 1.44 1.48 1.39 1.50

N.S. – – – – – –

⁄⁄⁄

18.95 18.46 18.08 18.48 17.12 16.84

11.45 12.66 14.06 13.28 11.13 12.92

339.39 304.84 305.82 440.45 432.17 427.09

266.97 276.69 278.14 325.21 268.4 279.5

81.85 93.89 93.18 106.88 82.14 95.16

Age category 21–29 30–39 40–49 50–64

N.S. – – – –

N.S. – – – –

⁄⁄

⁄⁄

⁄

⁄⁄

⁄⁄

⁄⁄

21.28 22.69 23.07 24.33

13.0 13.6 13.4 14.2

14.5 15.5 15.3 16.1

N.S. – – – –

⁄⁄⁄

258.6 275.6 284.3 299.2

323.23 366.25 386.58 423.78

286.47 290.94 265.06 287.47

86.80 89.27 87.14 105.53

TP⁄SGb

⁄⁄⁄

⁄⁄⁄

⁄

⁄

N.S.

N.S.

N.S.

⁄⁄

N.S.

N.S.

363.06 386.86

MLE = mouth level exposure, ADC = average daily consumption, based on all cigarettes smoked over 7 days, NEQ = nicotine equivalents, NNAL = 4-(methylnitrosamino)-1-(3pyridyl)-1-butanol, OH-cotinine = trans-30 -hydroxycotinine, N.S. = not significant (p > 0.05), ⁄p < 0.05, ⁄⁄p < 0.01, ⁄⁄⁄p < 0.001. a Smoker group: 8–10 mg group are those subjects who smoked a product of 8–10 mg ISO tar (not including the LSR product); 67 mg group are those smokers who smoked a product of 67 mg ISO tar. b TP * SG = Timepoint smoker group interaction. Timepoint and smoker group main effects were not removed from the model when the interaction of these parameters was significant, even if individually they were not statistically significant. c Consumption based on the number of spent filter-tips collected during the 24 h collection period from the cigarettes provided.

were lower than the 8–10 mg group at each timepoint, with the exception of timepoint 1 for MLE to nicotine per day. Daily exposure was lowest in the age group 21–29 compared to all other groups, and lower for the age groups 21–29 and 30–39 compared to 40–49 and 50–64. 3.3.2. Nicotine biomarkers of exposure The mean 24 h urinary NEQ value for the 67 mg group was statistically significantly lower than the 8–10 mg group and close to statistically significantly lower than the LSR group. The mean values of salivary cotinine and OH-cotinine for the 67 mg group were statistically significantly lower than the LSR and 8–10 mg groups. Whilst the effect of timepoint was significant for NEQ, no obvious trends were identified (Fig. 2a). Analysis by timepoint for salivary cotinine and OH-cotinine showed unusually high values at timepoint 4, with no obvious explanation for the observations (Fig. 2c and d). 3.3.3. Total NNAL per day Total NNAL per day showed lower mean values for the 67 mg group compared with the LSR and 8–10 mg groups. All smoker groups showed a step change at timepoint 4, although the extent of the increase was smaller for the 67 mg group compared with the LSR and 8–10 mg groups (Fig. 3). At timepoint 4, the mean value for the 67 mg group was significantly lower than the LSR group by 96.2 ng total NNAL/day (21%) and the 8–10 mg group by 130.2 ng total NNAL/day (26%). 3.4. Compensation index Calculation of the extent of compensatory smoking behaviour was assessed by considering changes in exposure endpoints on a per cigarette and per day basis for MLE estimates to tar and nicotine and for 24 h urinary NEQ. Table 4 shows compensation indices

calculated when a smoker switched from LSR to a product of 7 mg ISO tar or lower at adjacent timepoints. The median compensation index values were less than one but greater than zero for each measurement, indicating incomplete (partial) compensation. The degree of compensation was similar when calculated using the exposure endpoints of MLE to tar and nicotine per cigarette, 0.81 and 0.84, respectively. The degree of compensation increased marginally when calculated on a daily basis for MLE to tar and nicotine, 0.85 and 0.89, respectively. Compensation calculated using 24 h nicotine equivalents yielded a value of 0.78. Table 5 shows compensation indices calculated for those subjects whose remained smoking a product of 7 mg ISO or lower for one or more timepoints. Analysis of compensation indices by switch duration showed a statistically significant effect of switch duration for MLE to tar and nicotine on a per day basis. Contrast analysis revealed a downward trend with increasing switch duration, but was not statistically significant. There was no effect of switch duration on the compensation indices for MLE measures on a per cigarette basis or for NEQ. 4. Discussion This study was conducted to evaluate an approach for conducting long-term assessments of the patterns of cigarette consumption and exposure to cigarette smoke in a large population of adult smokers using biomarkers of tobacco smoke exposure and mouth level exposure estimates of tar and nicotine in their everyday environment. Whilst clinical confinement based studies provide more accurate, less variable data compared with nonresidential studies, some influence of participation in the study is likely, whereas non-residential based studies may provide data more representative of the smoker’s normal behaviour. As the objective of the study was to investigate smokers in their own

12

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

a

23

LSR ≤ 7 mg 8-10 mg 95% CI Upper Bound 95% CI Lower Bound

22

b

1.80

LSR ≤ 7 mg 8-10 mg 95% CI Upper Bound 95% CI Lower Bound

1.75 1.70 1.65

21

1.60 MLE Nicotine (mg/cig)

MLE Tar (mg/cig)

20

19

18

1.55 1.50 1.45 1.40 1.35

17

1.30 16

1.25 1.20

15

1.15 14

1.10 0

1

2

3 Time-point

4

5

6

0

1

2

3

4

5

6

Time-point

c

LSR ≤ 7 mg 8-10 mg 95% CI Upper Bound 95% CI Lower Bound

340

d

28

LSR ≤ 7 mg 8-10 mg 95% CI Upper Bound 95% CI Lower Bound

28 27 27 26

320

26

MLE Nicotine (mg/day)

MLE Tar (mg/day)

25 300

280

260

25 24 24 23 23 22 22 21 21 20

240

20 19 19

220

18 -1

0

1

2

3 Time-point

4

5

6

7

0

1

2

3 Time-point

4

5

6

Fig. 1. Mouth level exposure to tar/cigarette (1a), nicotine/cigarette (1b), tar/day (1c) and nicotine/day (1d) to reflect significant interaction of smoker group and timepoint. Data shown are LS means and 95% confidence intervals.

environment, nicotine equivalents and NNAL were chosen as the biomarkers of exposure as they are tobacco-specific and therefore not influenced by environmental and dietary sources (Hatsukami et al., 2006; Hecht, 2002). In addition, the use of the part-filter methodology has gained popularity recently and represents a non-invasive technique to provide estimates of MLE to nicotine and tar, in the subject’s own environment (Clayton et al., 2010; Ding et al., 2012; Pauly et al., 2009; Polzin et al., 2009). Mendes et al. (2008) aimed to address these concerns by inclusion of both a randomized, controlled, short-term phase and an unrestricted long-term follow-up phase in assessing the effects of forced switching. In that study, fewer subjects consented to continue in the long-term follow-up and for 2 of the 3 study groups, 50% of the subjects did not complete the study. In this study over 1000 smokers of the same 10 mg ISO tar product were recruited, a total of 546 subjects completed the final timepoint of the study, of which 462 subjects attended all seven timepoints. The higher than expected subject retention rate may be attributable to the non-confinement study design and the possibility for subjects to miss given timepoints. The provision of one day’s worth of cigarettes per timepoint in this study is unlikely to have influenced subject retention rates. Given that the reliability of self-reported cigarette consumption measures has been questioned (Mariner et al., 2011; Mendes et al., 2009) we measured average daily consumption from all cigarettes smoked across 7 days and the number of cigarettes smoked during the 24 h urine collection period (day 11). Cigarette consumption as measured by the two methods showed no significant variation

over the duration of the study or as a consequence of product switching. These findings are consistent with the trends observed by Guyatt et al. (1989), who studied a group of smokers for 5 months smoking their own cigarette product, followed by a forced switch to a cigarette with an ISO tar yield of at least 3 mg lower for 6 months. A longitudinal study of continuing smokers reported reductions in consumption over a 5 year period, though evidence of a survey effect was detected (Yong et al., 2012). However, increases in consumption have been reported in both clinical confinement (Shepperd et al., 2011) and non-confined study designs (Mendes et al., 2008; Sarkar et al., 2008). The provision of free cigarettes may have contributed to the results reported in those studies. In our study, subjects were provided with just one day’s worth of cigarettes per timepoint and thus we may have avoided this issue. In assessing the subjects’ exposure and consumption data throughout the study, each subject’s baseline data are considered in the linear mixed model, hence each subject acts as their own control. For MLE to tar per cigarette over the duration of the study, the effect of timepoint was statistically significant as was smoker group. Analysis of the interaction of smoker group and timepoint on the MLE estimates showed significant differences between the LSR and 67 mg groups at timepoints 1–3. This is presumably a result of switching to a cigarette of lower ISO tar yield, of at least 3 mg, as this difference was not observed in smokers who sideswitched to products of 8–10 mg ISO tar yield. A maximum difference of 3.04 mg tar/cig (16% relative to LSR group) was observed. Furthermore, differences were detected between the 67 mg and

13

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

a

16000

LS Mean 95% CI Upper Bound 95% CI Lower Bound

b

1550

LS Mean 95% CI Upper Bound 95% CI Lower Bound

1500

15000 1450

1400 Creatinine (mg/day)

Nicotine equivalents (ug/day)

14000

13000

12000

1350

1300

1250 11000 1200 10000 1150

9000

1100 0

1

2

3 Time-point

4

5

6

0

c

LS Mean 95% CI Upper Bound 95% CI Lower Bound

340

d

1

2

3 Time-point

4

5

6

120

LS Mean 95% CI Upper Bound 95% CI Lower Bound

115 110 105

OhCotinine (ng/ml)

Cotinine (ng/ml)

320

300

280

100 95 90 85 80

260

75 70

240 -1

0

1

2

3 Time-point

4

5

6

65

7

0

1

2

3 Time-point

4

5

6

Fig. 2. 24 h urinary nicotine equivalents (2a), 24 h urinary creatinine (2b), salivary cotinine (2c) and salivary hydroxycotinine (2d) to reflect significant effect of timepoint. Data shown are LS means and 95% confidence intervals.

LSR ≤ 7 mg 8-10 mg 95% CI Upper Bound 95% CI Lower Bound

600

550

NNAL (ng/day)

500

450

400

350

Table 4 Compensation indices following a switch from LSR to a product of 67 mg ISO tar at adjacent timepoints. Exposure endpoint

n

Compensation index, median (interquartile range Q1–Q3)

MLE MLE MLE MLE NEQ

99 99 99 99 98

0.81 0.84 0.85 0.89 0.78

tar/cigarette nicotine/cigarette tar/day nicotine/day

(0.45–1.17) (0.39–1.18) (0.31–1.47) (0.17–1.53) (1.54–1.92)

MLE = mouth level exposure, NEQ = nicotine equivalents.

300

250

200 0

1

2

3 Time-point

4

5

6

Fig. 3. 24 h urinary total NNAL to reflect significant interaction of smoker group and timepoint. Data shown are LS means and 95% confidence intervals.

8–10 mg groups at timepoints 1 and 2. However, these trends were not maintained for subsequent timepoints with reductions observed for the LSR and 8–10 mg groups at timepoints 5 and 6. Mouth level exposure to nicotine per cigarette data showed trends similar to MLE to tar. Significant differences were detected between the LSR and 67 mg groups at timepoints 1 and 3, with a maximum difference of 0.20 mg nicotine/cig (12% relative to LSR group). The downward trend observed for the LSR group’s MLE to

tar was less pronounced for MLE to nicotine and showed an increase at timepoint 6. The reason for the noticeable variation in MLE to tar and nicotine per cigarette for the LSR group is unknown and whilst analytical factors cannot be ruled out, it is unlikely as sample analysis was randomised across all subjects per timepoint. Over time, trends in daily MLE to tar and nicotine reflected the general trends observed for the per cigarette estimates. A significant difference was detected between the 67 mg and LSR groups at timepoint 3 of 54.1 mg tar/day (17.5 % relative to LSR group) and 3.57 mg nicotine/day (14.7% relative to LSR group). The additional variation introduced from the cigarette consumption data produced larger confidence intervals and in conjunction with the downward trend in MLE tar/cig for the LSR group, resulted in no other significant comparisons. High variability in urinary biomarkers of cigarette smoke exposure has been observed in studies with ambulatory collections

14

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

Table 5 Compensation indices for subjects who switched to and remained smoking a product of ISO tar yield of 7 mg or lower, for one or more timepoints. Exposure endpoint

MLE tar/cigarette MLE nicotine/cigarette MLE tar/day MLE nicotine/day NEQ Maximum observations, n

Compensation Index, median, by switch duration (months) 6

12

18

24

30

36

0.81 0.84 0.85 0.89 0.78 99

0.69 0.69 0.90 0.92 0.56 84

0.79 0.70 0.72 0.74 1.06 66

0.74 0.67 0.73 0.94 0.62 54

0.76 0.63 0.59 0.54 1.08 41

0.74 0.90 0.52 0.83 0.89 24

MLE = mouth level exposure, NEQ = nicotine equivalents.

(Mendes et al., 2008, 2009) and in this study we found that NEQ varied between timepoints, with no obvious trends. Whilst the extent of compliance in urine collection will affect variability, the trend in urinary creatinine (Fig. 2b) suggests that other sources of variation, such as individual’s metabolism rates for nicotine, contribute to the overall variation over time. Despite, the observed variation over time, a significant difference of 2.35 mg nicotine/day was observed in the 24 h urinary NEQ between the 67 mg and 8– 10 mg groups. The difference between the 67 mg and LSR group, whilst not statistically significant, was equivalent to 1.4 mg nicotine/day. The use of salivary biomarkers of exposure eliminates the concerns of sample integrity of 24 h ambulatory urine collections but is susceptible to the influence of proximity to recently smoked cigarettes, therefore leading to increased variability between smokers’ levels. The effect of smoker group closely matched that observed for NEQ, with lower values observed for the 67 mg group in comparison to the LSR and 8–10 mg groups for both cotinine and OH-cotinine. The values of cotinine and OH-cotinine were consistent over time, with the exception of distinctly high values at timepoint 4. The reason for the high salivary biomarker values at this timepoint is not understood and whilst no obvious analytical bias was identified, the fact that the increase occurred across all smoker groups indicates a common cause such as analytical accuracy or sample storage conditions. The inclusion of the urinary biomarker NNAL provides a biomarker of the particulate phase toxicant 4-(methylnitrosamino)1-(3-pyridyl)-1-butanone (NNK), (carcinogenic to humans; International Agency for Research on Cancer, Group 1). Total NNAL values were statistically lower for the 67 mg group compared with the LSR and 8–10 mg groups and increased at timepoint 4 across all smoker groups, although the extent of the increase was greatest for the LSR and 8–10 mg groups. The difference in the extent of the increase at timepoint 4 for the 67 mg resulted in a significant difference compared to the LSR and 8–10 mg groups. The possibility of an analytical bias was investigated by blinded repeat analysis of long-term stored samples from timepoints 3 and 4. The repeat analysis results confirmed the original values and implied that the analytical method was not the cause of the observed step change. Given that the actual cause of the step change in NNAL values remains unclear, the actual significance of the observed smoker group differences should be treated with caution. Further complication in the interpretation of total NNAL data arises due to the long half-life of 10–18 days (Goniewicz et al., 2009), which will result in contributions to the total NNAL measured from cigarettes smoked prior to the 24 h urine collection period. Whilst the NNK smoke yields were measured for each of the study samples, it was not feasible to obtain data on the products purchased and consumed by the subjects throughout the study. These potential differences in the NNK smoke yields between the study product batch and those purchased and consumed by the subjects may have contributed to the observed

trends in total NNAL and would suggest that measurement of NNAL in an non-confined study would require a study design which allows provision of a single batch of cigarettes of known NNK smoke yield. The secondary objectives of this study were to determine the extent of any compensatory smoking behaviour as a consequence of spontaneously switching to lower tar and nicotine yield cigarettes. Whilst compensation indices can be calculated when switching to higher yield products, subjects in this study could not switch to products with ISO tar yields greater than 10 mg due to the upper limit of 10 mg for cigarettes sold within the European Union (European Parliament and the Council of the European Union, 2001). Compensation indices were not calculated when subjects switched back to LSR from a product of 7 mg ISO tar or lower, due to the small number of observations (n = 4). The compensation indices (CiX) shown in Table 4 reflect incomplete (partial) compensation following a spontaneous switch from LSR to a cigarette of 7 mg ISO tar or lower. We are aware of only two published studies of spontaneous product switching (Lynch and Benowitz, 1987; Muhammad-Kah et al., 2011). Lynch and Benowitz reported data from 62 smokers who switched to products with a smoke machine-derived nicotine yield of at least 0.2 mg lower than their previous product of 3–6 years earlier. Daily consumption decreased on average by 6.6 cigarettes and plasma cotinine by 19%. However, cotinine data per cigarette revealed only slight downward changes, thus demonstrating that nicotine intake per cigarette was effectively maintained, i.e., complete compensation. In the study by Muhammad-Kah et al. (2011), 19 smokers had switched to products with 3 mg lower smoke machine-derived tar yields. Due to the limited sample size and variability of urinary biomarker levels, no firm conclusions were drawn regarding changes in exposure. The incomplete compensation findings from our study are in agreement with the conclusions made by Scherer and Lee (2014) based on a review of published forced switching studies. A comparison of the CiX values calculated using MLE estimates on a per cigarette and daily basis (Table 4) shows similar values and therefore implies that changes in cigarette consumption were minimal following spontaneous switching. This finding is consistent with the Scherer (1999) conclusion that compensation is driven by changes in smoking intensity rather than consumption. When considering the extent of compensation with increasing duration since the initial switch (Table 5), the CiX values derived from daily MLE values showed a downward, but non-statistically significant trend. These observations indicate that consumption may decrease somewhat over time for those smokers who remain smoking a cigarette of 7 mg ISO tar or lower. Variation in NEQ values observed in this study may explain the lack of any trend in CiX values with increasing switching duration and the value of 1.08 (which implies an increase in intake). The negative compensation index value observed for MLE nicotine/day implies that the reduction in intake is proportionally greater than the reduction in smoke yield, and in this case is likely to be the result of extreme changes in cigarette consumption for a small group size. In general the CiX values calculated from this study lie within the range of values observed from a number of forced-switching studies (Scherer and Lee, 2014) and suggest that forced-switching studies are a convenient study design to provide representative compensatory smoking behaviour data following a switch to a cigarette product of lower tar and nicotine yield. 5. Conclusions In undertaking the reported study we aimed to develop a suitable methodology for monitoring the smoking habits of a large smoker population in their everyday environment over an

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

extended period of time, whilst providing additional information on the effects of spontaneous product switching on smokers’ cigarette consumption and exposure to smoke constituents. Data from this study demonstrated that repeated measures of cigarette consumption remained stable and did not change as a consequence of either spontaneous side- or downswitching. MLE data for the smokers who did not switch product showed a downward trend part-way through the study. Whilst analytical variation may have caused or contributed to these trends, the results highlight the importance of minimising analytical variation and where possible limiting the variation of the study product over time. Statistically significant differences were detected between smoker groups, with smokers of products of 67 mg ISO tar obtaining lower yields compared with those who did not switch product and those who switched to products of 8–10 mg ISO tar. However, this trend was not observed for all timepoints. Biomarkers of exposure showed the greatest variation over time and whilst differences were detected between smoker groups, no clear trends were identified over time for the different smoker groups. Compensatory smoking behaviour as a consequence of spontaneous product switching was incomplete and in general agreement with findings from forced-switching studies. Changes in smoking intensity appear to be the driver of compensation rather than increases in cigarette consumption. In 2012, the US Food and Drug Administration issued draft guidance on the key areas to be investigated in making a modified risk tobacco product application, which includes the need to generate data on how consumers use a product both in controlled and natural environments and post-market studies to provide longer-term assessment of exposure and health outcomes (US Food and Drug Administration, 2012). Data from this study illustrate some of the challenges in measuring smokers’ long-term exposure to smoke constituents and whilst non-confined study designs aim to reflect natural behaviours, the variability of biomarkers may limit their value. Conflict of interest This work was funded by British American Tobacco (BAT), and all authors are full time employees of BAT. Acknowledgments The authors would like to thank Madeleine Ashley and Jon Sheppard for their assistance in preparing the study data, and Derek Mariner and Christopher Proctor for their assistance with the manuscript. We also thank Harrison Clinical Research Deutschland GmbH and Ipsos Observer for conducting the study and ABF GmbH Munich, Germany for bioanalytical analysis.

References Ashley, M., Saunders, P., Mullard, G., Prasad, K., Mariner, D., Williamson, J., 2011. Smoking intensity before and after introduction of the public place smoking ban in Scotland. Regul. Toxicol. Pharmacol. 61, S60–S65. Benowitz, N.L., 2001. Compensatory smoking of low-yield cigarettes. In: Public Health Services, National Institutes of Health, National Cancer Institute (Eds.), Smoking and Tobacco Control. Risks Associated with Smoking Cigarettes with Low Machine-Measured Yields of Tar and Nicotine. NIH Publication, Bethesda, MD, pp. 39–63. Benowitz, N.L., Jacob III, P., Bernert, J.T., Wilson, M., Wang, L., Allen, F., Dempsey, 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.

15

Blaszkewicz, M., Liesenhoff-Henze, K., 2012. Creatinine in urine [Biomonitoring Methods, 2010]. The MAK Collection for Occupational Health and Safety. Wiley, pp. 169–184. Burns, D.M., Dybing, E., Gray, N., Hecht, S., Anderson, C., Sanner, T., O’Connor, R., Djordjevic, M., Dresler, C., Hainaut, P., Jarvis, M., Opperhuizen, A., Straif, K., 2008. Mandated lowering of toxicants in cigarette smoke: a description of the World Health Organization TobReg proposal. Tob. Control 17, 132–141. 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 139, 291–299. Clayton, P.M., Cunningham, A., van Heemst, J.D.H., 2010. Quantification of four tobacco-specific nitrosamines in cigarette filter tips using liquid chromatography–tandem mass spectrometry. Anal. Meth. 2, 1085–1094. Cunningham, A., Sommarström, J., Sisodiya, A.S., Errington, G., Prasad, K., 2014. Longitudinal study of long-term smoking behaviour by biomarker-supported determination of exposure to smoke. BMC Public Health 14, 348. Ding, Y.S., Chou, T., Abdul-Salaam, S., Hearn, B., Watson, C.H., 2012. Development of a method to estimate mouth-level benzo[a]pyrene intake by filter analysis. Cancer Epidemiol. Biomarkers Prev. 21, 39–44. European Parliament and the Council of the European Union, 2001. Directive 2001/ 37/EC of the European Parliament and of the Council of 5 June 2001 on the approximation of the laws, regulations and administrative provisions of the Member States concerning the manufacture, presentation and sale of tobacco products. Off. J. Eur. Communities 194, 26–34. Federal Trade Commission, 1967. FTC to begin cigarette testing. [http://www.pmi. com/eng/tobacco_regulation/regulating_tobacco/regulation_of_tobacco_smoke/ documents/1967_ftc_press_release_ftc%20to%20begin%20cigarette%20testing. pdf]. (accessed December 3, 2014). Frost, C., Fullerton, F.M., Stephen, A.M., Stone, R., Nicolaides-Bouman, A., Densem, J., Wald, N.J., Semmence, A., 1995. The tar reduction study: randomised trial of the effect of cigarette tar yield reduction on compensatory smoking. Thorax 50, 1038–1043. Goniewicz, M.L., Havel, C.M., Peng, M.W., Jacob III, P., Dempsey, D., Yu, L., ZielinskaDanch, W., Koszowski, B., Czogala, J., Sobczak, A., Benowitz, N.L., 2009. Elimination kinetics of the tobacco-specific biomarker and lung carcinogen 4(methylnitrosamino)-1-(3-pyridyl)-1-butanol. Cancer Epidemiol. Biomarkers Prev. 18, 3421–3425. Guyatt, A.R., Kirkham, A.J.T., Mariner, D.C., Baldry, A.G., Cumming, G., 1989. Longterm effects of switching to cigarettes with lower tar and nicotine yields. Psychopharmacology 99, 80–86. Hammond, D., Fong, G.T., Cummings, K.M., Hyland, A., 2005. Smoking topography, brand switching, and nicotine delivery: results from an in vivo study. Cancer Epidemiol. Biomarkers Prev. 14, 1370–1375. Hatsukami, D.K., Benowitz, N.L., Rennard, S.I., Oncken, C., Hecht, S.S., 2006. Biomarkers to assess the utility of potential reduced exposure tobacco products. Nicotine Tob. Res. 8, 600–622. Hecht, S.S., 2002. Human urinary carcinogen metabolites: biomarkers for investigating tobacco and cancer. Carcinogenesis 23, 907–922. Hecht, S.S., Murphy, S.E., Carmella, S.G., Li, S., Jenson, J., Le, C., Joseph, A.M., Hatsukami, D.K., 2005. Similar uptake of lung carcinogens by smokers of regular, light, and ultralight cigarette. Cancer Epidemiol. Biomarkers Prev. 14, 693–698. International Agency for Research on Cancer, 2004. Tobacco smoke and involuntary smoking, Volume 83. Lyon, France: IARC (IARC monographs on the evaluation of carcinogenic risks to humans). [http://monographs.iarc.fr/ENG/Monographs/ vol83/mono83.pdf] (accessed January 5, 2015). International Conference on Harmonisation, 1996. ICH Tripartite Guidelines for Good Clinical Practice, Adopted by CPMP, issued as CPMP/ICH/135/95. International Conference on Harmonisation, Geneva. International Organization for Standardization, 2000. Routine analytical cigarettesmoking machine – definitions and standard conditions. ISO 3308:2000. Geneva: International Organization for Standardization. Lynch, C.J., Benowitz, N.L., 1987. Spontaneous cigarette brand switching: consequences for nicotine and carbon monoxide exposure. Am. J. Public Health 78, 1191–1194. Mariner, D.C., Ashley, M., Shepperd, C.J., Mullard, G., Dixon, M., 2011. Mouth level smoke exposure using analysis of filters from smoked cigarettes: a study of eight countries. Regul. Toxicol. Pharmacol. 61, S39–S50. Meger, M., Meger-Kossien, I., Schuler-Metz, A., Janket, D., Scherer, G., 2002. Simultaneous determination of nicotine and eight nicotine metabolites in urine of smokers using liquid chromatography–tandem mass spectrometry. J. Chromatogr. B 778, 251–261. 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, 295–305. Mendes, P., Liang, Q., Frost-Pineda, K., Munjal, S., Walk, R.A., Roethig, H.J., 2009. The relationship between smoking machine derived tar yields and biomarkers of exposure in adult cigarette smokers in the US. Regul. Toxicol. Pharmacol. 55, 17–27. Muhammad-Kah, R.S., Mendes, P., Rimmer, L., Liang, Q., Serafin, R., Roethig, H.J., Sarkar, M., 2011. Exposure to cigarette smoke constituents in a population of adult cigarette smokers in the U.S. who spontaneously switched to cigarettes with lower or higher machine measured ‘tar’ yield. Beiträge zur Tabakforschung Int 24, 166–173.

16

A. Cunningham et al. / Regulatory Toxicology and Pharmacology 72 (2015) 8–16

Pauly, J.L., O’Connor, R.J., Paszkiewicz, G.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, 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 exposure to select mainstream smoke constituents from discarded cigarette filter butts. Nicotine Tob. Res. 11, 868–874. Sarkar, M., Kapur, S., Frost-Pineda, K., Feng, S., Wang, J., Liang, Q., Roethig, H., 2008. Evaluation of biomarkers of exposure to selected cigarette smoke constituents in adult smokers switched to carbon-filtered cigarettes in short-term and longterm clinical studies. Nicotine Tob. Res. 10, 1761–1772. Scherer, G., 1999. Smoking behaviour and compensation: a review of the literature. Psychopharmacology 145, 1–20. 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. Scherer, G., Lee, P.N., 2014. Smoking behaviour and compensation: a review of the literature with meta-analysis. Regul. Toxicol. Pharmacol. 70, 615–628. Shepperd, C.J., Eldridge, A.C., Errington, G., Dixon, M., 2011. 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. Regul. Toxicol. Pharmacol. 61, S13–24. St. Charles, F.K., Ashley, M., Shepperd, C.J., Clayton, P., Errington, G., 2009. A robust method for estimating human smoked cigarette yields from filter analysis data. Beiträge zur Tabakforschung Int 23, 232–243. Stratton, K., Shetty, P., Wallace, R., Bondurant, S. (Eds.), 2001. Clearing the smoke. Assessing the science base for tobacco harm reduction. The Institute of Medicine, National Academy of Sciences. National Academy Press, Washington.

U.S. Department of Health and Human Services, 1982. The health consequences of smoking: cancer: a report of the Surgeon General. US Department of Health and Human Services, Public Health Service, Office on Smoking and Health, Washington, DC, DHHS Publication No (PHS) 82–50179. U.S. Department of Health and Human Services, 1983. The health consequences of smoking: cardiovascular disease: a report of the Surgeon General. US Department of Health and Human Services, Public Health Service, Office on Smoking and Health, Rockville, MD, DHHS Publication No. (PHS) 84–50204. U.S. Department of Health and Human Services, 1984. The health consequences of smoking: chronic obstructive lung disease: a report of the Surgeon General. US Department of Health and Human Services, Public Health Service, Office on Smoking and Health, Rockville, MD, DHHS Publication No. (PHS) 84–50205. U.S. Department of Health and Human Services, 2004. The health consequences of smoking: a report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, Washington, D.C. U.S. Food and Drug Administration 2012. Modified Risk Tobacco Product Applications – Draft Guidance for Industry. [http://www.fda.gov/downloads/ TobaccoProducts/GuidanceComplianceRegulatoryInformation/UCM297751. pdf] (accessed December 3, 2014). World Health Organization, 2007. The scientific basis of tobacco product regulation: report of a WHO study Group. Geneva: World Health Organization. Geneva, WHO Technical Report Series 945, 1–112. World Medical Association, 1996. Declaration of Helsinki – ethical principles for medical research involving human subjects. World Medical Association, FerneyVoltaire. Yong, H.-H., Borland, R., Thrasher, J.F., Thompson, M.E., 2012. Stability of cigarette consumption over time among continuing smokers: a latent growth curve analysis. Nicotine Tob. Res. 14, 531–539.