Assessment of dioxin and dioxin-like compounds in mainstream smoke from selected US cigarette brands and reference cigarettes

Available online at www.sciencedirect.com

Food and Chemical Toxicology 46 (2008) 1721–1733 www.elsevier.com/locate/foodchemtox

Assessment of dioxin and dioxin-like compounds in mainstream smoke from selected US cigarette brands and reference cigarettes C.L. Wilson *, J.A. Bodnar, B.G. Brown, W.T. Morgan, R.J. Potts, M.F. Borgerding Research and Development, R.J. Reynolds Tobacco Company, Winston-Salem, P.O. Box 1487, NC 27102, United States Received 7 May 2007; accepted 7 January 2008

Abstract Mainstream cigarette smoke (MSS) from 12 US cigarette brands and two reference cigarettes was evaluated to determine concentrations of dioxins (i.e., polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and dioxin-like polychlorinated biphenyls (PCBs)). The study included three ‘tar’ ranges based on Federal Trade Commission (FTC) determination: Low Yield (LY) 6 5.5, Medium Yield (MY) 9.6–12.2, and High Yield (HY) P 14.5 mg/cig. Of the brands studied, the HY cigarettes yielded the greatest mean concentrations of 2005 World Health Organization Toxic Equivalents (WHO-TEQs) on a per cigarette basis. WHOTEQ levels in LY cigarettes were significantly lower than for HY cigarettes (p = 0.039) on a yield per cigarette basis and WHO-TEQ concentrations correlated with ‘tar’ yield (r = 0.73, p = 0.007), as did concentration on a WHO-TEQ per body mass per day basis (r = 0.73, p = 0.007). However, a statistically significant relationship was not observed between ‘tar’ yield levels and WHO-TEQ concentrations on a per mg Total Particulate Matter (TPM) basis. Concentrations for all brands tested ranged from 0.44 to 3.88 fg WHO-TEQ/ mg TPM. Maximum daily exposure estimates calculated from this range (0.004–0.074 pg WHO-TEQ/kg bw/day) are below the current WHO Tolerable Daily Intake range of 1–5 pg/kg bw/day. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: PCDD; PCDF; Mainstream smoke; PCB; Risk assessment; Cigarette

1. Introduction Mainstream cigarette smoke (MSS) consists of over 4700 constituents that exist in a dynamic and chemically complex aerosol and can be categorized as existing primarily in either a gas phase or a gas suspended particulate Abbreviations: EMPC, estimated maximum possible concentration; FTC, federal trade commission; HAH, halogenated aromatic hydrocarbons; HY, high yield; ISO, International Organization for Standardization; LOD, limit of detection; LOQ, limit of quantitation; LY, low yield; MSS, mainstream cigarette smoke; MY, medium yield; PAHs, polynuclear aromatic hydrocarbons; PCB, polychlorinated biphenyl; PCDD, polychlorinated dibenzo-p-dioxin; PCDF, polychlorinated dibenzofuran; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin; TEF, toxic equivalency factor; TEQ, toxic equivalent; TPM, total particulate matter; WHO-TEQ, World Health Organization toxic equivalent. * Corresponding author. Tel.: +1 336 741 5466; fax: +1 336 741 0815. E-mail address: [email protected] (C.L. Wilson). 0278-6915/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.fct.2008.01.009

phase (Dube and Green, 1982). The gas phase of a non-filter cigarette consists of nearly 500 individual volatile compounds, including carbon monoxide, nitrogen oxides, and ammonia and comprises roughly 95% of the weight of MSS (Hoffmann and Hoffmann, 1997). The particulate phase contains more than 3500 semivolatile and nonvolatile individual compounds, including nicotine, polynuclear aromatic hydrocarbons (PAHs) (Hoffmann and Hoffmann, 1997) and halogenated aromatic hydrocarbons (HAHs) (Muto and Takizawa, 1989; Ball et al., 1990; Lo¨froth and Zebhr, 1992; Smith et al., 2004). Composition of both gas and particulate phases vary with a broad range of cigarette design features (Borgerding and Klus, 2005). MSS typically is analyzed for both yield and composition. Yield measurements include the determination of ‘tar’, nicotine, and carbon monoxide generated under standard conditions defined by domestic and international

1722

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

regulatory bodies (i.e., US Federal Trade Commission (FTC), International Organization for Standardization (ISO)). MSS ‘tar’ is a chemically complex mixture defined as the Total Particulate Matter (TPM) in smoke less the weight of nicotine and water (Pillsbury et al., 1969). The ‘tar’ range in the worldwide marketplace is broad. The products studied in the present report represent that broad range through three ‘tar’ yield categories: Low Yield (LY), Medium Yield (MY), and High Yield (HY) delivering 1.0– 5.5 mg, 9.6–12.2 mg, and 14.5–16.8 mg ‘tar’ per cigarette, respectively. A range of design features contribute to the ‘tar’ yield of a cigarette, including tobacco type, tobacco weight, filter composition, filter ventilation, structural dimensions, and papers (reviewed by Norman, 1999). Methods to assess cigarette ‘tar’ yield have been standardized for regulatory and research and development purposes. These national and international protocols provide a standardized basis for comparing cigarette MSS yields. Under the FTC machine smoking regimen, applicable for determining cigarette yield ratings for comparison of products sold in the United States, a 35 ml puff of 2s duration is taken every 60 s (i.e., ‘‘35/60/2”). Fundamentally similar to the FTC method with regard to smoking regimen, ISO Method 4387 provides further guidance on cigarette conditioning parameters and product sampling procedures. A more intense smoking regimen used in the present study, 60/30/2, is designed to estimate smoke yields under more stringent smoking conditions. It should be noted that, while MSS yields as measured by an intense machine smoking regimen may estimate maximum potential smoke exposures, actual exposure to MSS constituents is highly variable and driven by a broad range of individual smoking behaviors. Smoke composition investigations have focused on a range of applications, including regulatory, product devel-

Table 1 Features of the 12 cigarette brands studied Cigarette Identification

Cigarette Description

FTC ‘tar’ (mg/cig) (35/60/2)

Length (mm)

Average TPM (mg/cig) (60/30/2)

76 78 79 80 81 85 86 87 88 89 91 92 K2R4F K1R5F

LY MY HY LY LY MY MY MY HY HY HY LY MY LY

5.5 10.3 14.5 5.0 5.0 10.5 9.6 12.2 14.5 16.8 14.5 1.0 11.6 2.0

83 83 83 83 83 83 100 120 83 100 83 100 83.9 83.9

21.2 31.8 50.8 18.5 20.1 37.9 32.8 40.1 62.1 56.9 61.0 6.7 31.7 8.9

LY: Low ‘tar’ Yield; MY: Medium ‘tar’ Yield; HY: High ‘tar’ Yield. LY, MY, and HY specify three ‘tar’ ranges based on the FTC machine smoking regimen. TPM: Total Particulate Matter.

opment, and health risk assessment purposes. With regard to the latter, several compilations of chemicals present in cigarette smoke as potential toxicants have been published (Hoffmann and Hoffmann, 1997; Rodgman and Green, 2003). Qualitative and quantitative investigations of specific classes of toxicants, such as the PAHs, have also been published (Rodgman, 2001). Interestingly, the individual concentrations for many analytes bear a relatively strong positive correlation with MSS ‘tar’ yield (Chepiga et al., 2000; Borgerding and Klus, 2005). The dioxins are members of a broader group of HAHs, which includes PCDDs, PCDFs, PCBs, and others. The term ‘‘dioxin” and ‘‘dioxin-like” includes the PCDDs, the 2,3,7,8-substituted PCDFs, and certain specific PCBs. These compounds are collectively called ‘‘dioxin” or ‘‘dioxin-like” because they induce a common pattern of toxic responses, exemplified by the family’s prototypical and most potent congener, 2,3,7,8-tetrachlorodibenzo-pdioxin (TCDD). Dioxins are highly lipophilic, ubiquitous contaminants of the global ecosystem and have been described in virtually every component of the biosphere, including air, aquatic sediment, fish, and wildlife and human adipose tissue, milk, and blood (Ballschmiter et al., 1980; DeVoogt and Brinkman, 1989; Brinkman and DeKok, 1989; Rappe et al., 1979; Schecter et al., 2006). Invariably, these compounds exist as complex mixtures of isomers and congeners in a diversity of analytes, complicating risk and hazard assessment efforts. The dioxins induce a common spectrum of biochemical and toxic effects through a receptor-mediated mechanism of action, facilitating development and application of a relative potency factor risk assessment approach. Application of a Toxic Equivalents (TEQs) approach relies on several assumptions, with the most basic being that the combined effects of the different congeners are additive, although mixtures of PCDDs/PCDFs that also contain certain PCB congeners (e.g., PCBs 77 and 153) exhibit antagonistic interactive responses (Safe, 1992, 1998). In addition, it is assumed that all the individual compounds act through the same biologic or toxic mode of action and that doseresponse curves for the different congeners are parallel (Safe, 1998). The TEQ approach is a scheme used to express the toxicity of an individual dioxin relative to that of TCDD. The overall potency or TEQ of a mixture is defined by the following equation: TEQ ¼ R½C i   TEFi where Ci is the concentration of an individual congener and TEFi is the Toxic Equivalence Factor (TEF) of an individual congener. Individual TEF values have been revised by multiple organizations over the last decades; however, a 2005 World Health Organization (WHO) re-evaluation affirmed the plausibility and feasibility of the TEF approach for risk assessment of chemicals with dioxin-like properties (Van den Berg et al., 2006).

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

Cigarette smoke has been recognized as a potential lowlevel emission source of dioxins since first reported by Bumb et al. (1980) and this is consistent with reports that dioxins arise from other organic material combustion processes (USEPA, 2000). Subsequent studies have varied with regard to smoking regimen used, sample collection, and analytical methods. However, most investigations of dioxins in cigarette smoke report the higher chlorinated CDD/CDF congeners (i.e., hexa-, hepta- and octa-CDDs) predominating (Muto and Takizawa, 1989; Ball et al., 1990; Lo¨froth and Zebhr, 1992; Smith et al., 2004; Aoyama et al., 2003). Herein, we report MSS analysis of dioxins from 12 US cigarette brands produced by the same manufacturer and two University of Kentucky Reference Cigarettes, K1R5F and K2R4F. A previous report from our institution based on a mixture of brands, including brands with charcoal filters, sold in several countries suggested an inverse relationship between cigarette TPM and WHOTEQ concentration (Smith et al., 2004). In contrast, results from the present study of a range of brands from the US market representative of all typical design features in the market demonstrated that HY cigarettes yielded the greatest mean concentrations of WHO-TEQs (hereinafter referred to as TEQs) on a per cigarette basis. TEQ levels in MSS from LY cigarettes were significantly lower than for HY cigarettes on a yield per cigarette basis and TEQ concentrations correlated with ‘tar’ yield, as did concentration on a TEQ per body mass per day basis. Importantly, however, a statistically significant relationship was not observed between ‘tar’ yield levels and TEQ concentrations on a per mg TPM basis. Finally, calculated maximum daily exposure estimates are below the current WHO Tolerable Daily Intake range. 2. Materials and methods 2.1. Cigarette descriptions All test cigarettes were obtained from a single US manufacturer and had design features typical of cigarettes common to the US market today, including representative ranges of tobacco blend, tobacco weight, filter type, filter ventilation, paper porosities, smoke yields, and paper types. With regard to FTC ‘tar’ range, four brands were chosen from each of the LY, MY, and HY categories. All cigarettes used in the study were manufactured in 2005 except for the 1R5F and 2R4F University of Kentucky Reference Cigarettes, manufactured in 1989 and 2005, respectively. None of the cigarettes contained charcoal in the filter section. Table 1 describes the characteristics of the 12 cigarette brands and two reference cigarettes used in this study.

2.2. FTC tar determination For determination of FTC ‘tar’ yield, cigarettes were smoked on a 20port linear smoking machine using the FTC smoking regime. The FTC smoking regime consists of a 35 ml puff taken every 60 s for a 2-s duration (35/60/2) with no ventilation hole blocking.

2.3. Collection of total particulate matter All machine smoking procedures were carried out in an ISO 17025 accredited laboratory. For collection of TPM, cigarettes were smoked by

1723

the RJRT Analytical Chemistry Laboratories on Borgwaldt rotary smoking machines using an intense smoking regimen (60 ml puff taken every 30 s for a 2s duration (60/30/2) with no cigarette ventilation holeblocking). Varying numbers of cigarettes were smoked for each brand in order to collect about 150 mg of TPM on each Cambridge filter. Four Cambridge filters per brand were placed in a collection vial. The vials were sealed and held in a freezer until delivery to the contract analytical laboratory for analysis. Triplicate samples were prepared for each brand.

2.4. Analytical method Fourteen sets of filter pads were analyzed in triplicate from the mechanical smoking of 12 cigarette brands and two reference cigarettes. Three sets of blank filter pads were also analyzed to determine the analyte contribution from the filter pads. The filters were analyzed under contract by Eno River Labs (Durham, NC) for the 17 2,3,7,8-substituted PCDDs and PCDFs and for the 12 non-ortho and mono-ortho substituted PCBs for which TEFs have been assigned by WHO (Van den Berg et al., 2006). While two EPA methods (1613 and 8290) are available for dioxin analysis, EPA Method 8290 was used for PCDD and PCDF analysis to provide a more reasonable basis for comparison to previous studies (Smith et al., 2004). As the functional differences between the two methods are minimally significant, the two methods would be expected to yield comparable results. The lowest sample specific detection limit for each congener is listed in Table 2. Filter pads were spiked with an internal standard and particulate matter was extracted using toluene as an extraction solvent as recommended by Method 8290 for solid matrices. The extracts were concentrated, and interferences to the analysis were removed using a multi-step extract cleanup process. Briefly, sample cleanup included reaction with

Table 2 Lowest sample specific detection limits Analyte

Lowest sample specific detection limit (pg/g TPM)

2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF PCB81 PCB77 PCB123 PCB118 PCB114 PCB105 PCB126 PCB167 PCB156 PCB157 PCB169 PCB189

0.8 0.8 0.7 0.7 0.6 2.6 6.1 0.5 0.7 0.6 0.5 0.4 0.5 0.5 0.7 0.8 3.7 17.2 15.8 21.8 259 17.5 75.4 19.2 38.7 46.8 19.9 20.7 15.2

Results are presented as pg congener per gram Total Particulate Matter (TPM).

1724

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

sulfuric acid on silica gel to remove bulk interferences, chromatography on deactivated alumina to remove phenols, chromatography on silica gel coated with AgNO3 to remove PAHs, and chromatography on activated acidic alumina to remove PCBs and PCDE interferences. The extracts were taken to dryness, reconstituted in a recovery standard and analyzed by isotope dilution gas chromatography-high resolution mass spectrometry. As this sample workup procedure had sufficient capacity to handle the amount of ‘tar’, no adjustments were needed to adjust for ‘tar’ quantity relative to the recommended sample quantities within the Method. Analysis of PCBs was performed using EPA Method 1668A with the following enhancements: perfluorotributylamine was used as a tuning standard; all channels except M  70, M + 35, and M + 70 were monitored; quantitation limits were based upon calibration; and, the modified method uses a DB-35MS column. A laboratory blank and a positive control sample were extracted and analyzed with each sample batch to monitor and demonstrate laboratory control.

2.5. Data reporting and statistical methods For congeners with analytical values below the Limit of Detection (LOD), we assumed that levels for nondetected congeners were zero. A similar approach has been used in calculating TEQs from data sets containing multiple values below the LOD (Kiviranta et al., 2004; Smith et al., 2004; Charnley and Doull, 2005), and this approach is consistent with US Food and Drug Administration guidance stating that setting non-detects equal to zero (i.e., including only values actually measured) provides more realistic intake estimates (USFDA, 2004). Individual values reported with a J-qualifier (Estimated Value – Below the Calibration Range) were used as reported. Concentration data were log-transformed to make within-cigarette variation more consistent and two approaches were used to evaluate the relationship between cigarette ‘tar’ yield and logs of TEQ concentration. The first approach used analysis of variance to compare the means of logs of concentration for HY, MY, and LY styles. Means for styles from each ‘tar’ category were compared to each other, with p-values adjusted using the Bonferroni method. Reported ‘tar’ category mean values were backtransformed to original units for simplicity of presentation. The second approach calculated Pearson correlation coefficients for dioxin concentrations vs. ‘tar’ yield. This analysis was applied to mean TEQ concentrations for each cigarette. Again, p < 0.05 was required for statistical significance.

Although detectable, dioxin concentrations in most samples were extremely low. The prototypical compound in the series, 2,3,7,8-TCDD, was not quantifiable in any of the samples, but did meet criteria to be included in calculations of one sample as an EMPC. In general, PCDFs were more prevalent in the samples than PCDDs, but the lower chlorinated analytes were below the LOQ in all cases. Most of the samples showed concentrations of 1,2,3,5,6,7,8-HpCDD and OCDD, and to a lesser extent, OCDF above background levels (Tables 3A–3C). These higher chlorinated congeners are stable and solely particle bound and are the most commonly observed PCDD/ PCDFs in ambient samples. However, their toxicity is also the lowest among the 2,3,7,8-substituted analytes (Van den Berg et al., 2006). The two PCBs analyzed with the highest TEFs, PCB126 and PCB169, were not detected in any of the samples. The congeners most commonly detected were PCB118 and PCB105, two PCBs found ubiquitously in the environment. The PCB analysis showed lower sensitivity, in general, leading to higher LODs and consequently higher TEQ values. This is likely a result of organic components of the extract not being completely removed in the sample cleanup. The results in Fig. 1 show the relative contribution to TEQs for each ‘tar’ category. For each of the three ‘tar’ categories, PCDF congeners dominated the mixture with or without inclusion of EMPCs. However, the lower chlorinated PCDF congeners, in particular, drove the observed pattern, but only when EMPCs were used in the calculations. PCDD congeners comprised roughly 12–27% of the mixture analyzed for the three ‘tar’ categories, with 1,2,3,4,6,7,8-HpCDD and OCDD being the major PCDD congeners detected. 2,3,7,8-TCDD was detected in only one replicate of one brand, Brand 79, but the analytical value was categorized as an EMPC. PCB 118 and PCB 105 dominated the PCB contribution to the TEQ of the mixtures.

3. Results 1

Relative Contribution to TEQ (fg TEQ/mg TPM)

Analytical results for dioxin concentrations below the Limit of Quantitation (LOQ) but above the LOD frequently are reported as Estimated Maximum Possible Concentrations (EMPC). As defined in the US EPA SW-846 guidelines for Method 8280A (USEPA, 1996), an EMPC is calculated for values that meet all validation criteria except that ion ratios exceed the prescribed limits. Further guidance suggests against the use of EMPC values in TEQ calculations (USEPA, 1996). As EMPC could be considered a ‘‘worst case” estimate of concentration and potential exposure, we included EMPC values in estimating potential maximum exposure, and excluded EMPC values (i.e., designated as nil) in TEQ calculations when calculating results on a basis more consistent with previous reports and guidance (USEPA, 1996; Kiviranta et al., 2004). Upon comparing the results, we found no difference between the two approaches in final outcomes of risk calculations or conclusions.

0.9 0.8 0.7 0.6

PCB PCDF PCDD

0.5 0.4 0.3 0.2 0.1 0 LY

MY

HY

2R4F

1R5F

'Tar' Category

Fig. 1. Relative contribution of PCDDs, PCDFs, and PCBs to TEQ in cigarette MSS collected from cigarettes of three different ‘tar’ categories. TPM, Total Particulate Matter; LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

4.00 3.00 2.00 1.00

†

*

0.00 MY

HY

2R4F

1R5F

100.0

†

80.0

*

60.0 40.0 20.0

*

†

0.0

HY

LY

MY

HY

2R4F

1R5F

'Tar' Category

5.00 4.00 3.00 2.00 1.00 0.00 76 80

81

92

78

85

86

87

88

91

79

89 2R4F 1R5F

Brand

Fig. 2A. Individual brand differences in Toxic Equivalents (TEQ) when normalized to Total Particulate Matter (TPM). Shaded bars represent results obtained using Estimated Maximum Permissible Concentrations (EMPCs). Information on each cigarette brand can be found in Table 1. LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

200.0

LY fg TEQ per cig

5.00

Fig. 3A. ‘Tar’ category differences in Toxic Equivalents (TEQ) when normalized to Total Particulate Matter (TPM). Shaded bars represent results obtained using Estimated Maximum Permissible Concentrations (EMPCs). *, significant difference between LY and 1R5F reference cigarette (p < 0.05);  , significant difference between 2R4F and 1R5F reference cigarettes (p < 0.05). LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

MY

HY

Fig. 3B. ‘Tar’ category differences in Toxic Equivalents (TEQ) when normalized on a per cigarette basis. Shaded bars represent results obtained using Estimated Maximum Permissible Concentrations (EMPCs).  , significant difference between LY and HY categories (p < 0.05); *, significant difference between 2R4F and 1R5F reference cigarettes (p < 0.05). LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

0.050

pg TEQ / kg bw / d

fg TEQ per mg TPM

6.00

MY

*†

6.00

'Tar' Category

7.00

LY

1725

7.00

LY

fg TEQ per cig

Results for the 2R4F reference cigarette revealed OCDD and 1,2,3,4,6,7,8-HpCDD dominating the congener pattern, with PCDDs contributing over 95% of the relative TEQ value. PCB 118 contributed to <5% of the overall response in the 2R4F cigarette. The relative contribution to the total TEQ yield from PCDDs, PCDFs, and PCBs in the 1R5F reference cigarettes was 49%, 41%, and 10%, respectively. PCDD congeners in the 1R5F reference cigarettes arose from 1,2,3,4,6,7,8-HpCDD, 1,2,3,6,7,8HxCDD, and 1,2,3,7,8,9-HxCDD, while 2,3,4,7,8-PeCDD was the major contributor to the PCDF fraction. TCDF was detected in a single sample, but the analytical value was categorized as an EMPC. Among the PCBs in the 1R5F cigarette, PCBs 105 and 118 were the congeners detected in greatest quantity. The results in Fig. 2 show the individual brand mean TEQs in MSS of the 12 brands and two reference cigarettes studied. On a per mg TPM basis (Fig. 2A) and a per cigarette basis (Fig. 2B), brands were largely consistent within their respective ‘tar’ categories with the exception of Brand 85, a MY cigarette, and Brand 91, a HY cigarette. Likewise, TEQ values for most of the brands tested and both reference cigarettes are consistent with regard to inclusion

fg TEQ per mg TPM

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

†

0.040

*

0.030 0.020 0.010

*

†

0.000

150.0

LY

MY

HY

2R4F

1R5F

'Tar' Category

100.0 50.0 0.0 76

80

81

92

78

85

86

87

88

91

79

89 2R4F 1R5F

Brand

Fig. 2B. Individual brand differences in Toxic Equivalents (TEQ) when normalized to number of cigarettes. Shaded bars represent results obtained using Estimated Maximum Permissible Concentrations (EMPCs). Information on each cigarette brand can be found in Table 1. LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

Fig. 4. Daily exposure estimates of three ‘tar’ categories. Shaded bars represent results obtained using Estimated Maximum Permissible Concentrations (EMPCs). Exposure point estimates were calculated by multiplying fg TEQ/cig by the measured daily cigarette consumption of 38 cigs/day and dividing by a 75 kg body weight. Body weight (median body weight of 75 kg) and daily cigarette consumption (95th percentile 38.181 cig/day) were obtained from the 1999 to 2000 NHANES survey data (National Center for Health Statistics, 1999).  , significant difference between LY and HY categories (p < 0.05); *, significant difference between 2R4F and 1R5F reference cigarettes (p < 0.05). LY, low ‘tar’ yield; MY, medium ‘tar’ yield; HY, high ‘tar’ yield; 2R4F, University of Kentucky medium ‘tar’ yield reference cigarette; 1R5F, University of Kentucky low ‘tar’ yield reference cigarette.

1726

Table 3A Analytical results for Low Yield cigarette brands Analyte

TEF

Cigarette identification (Low Yield) 76

1 1 0.1 0.1 0.1 0.01 0.0003 0.10 0.03

ND ND ND ND ND 290 (45) 1077 (280) 217 (48) 47 (82)

2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF

0.30 0.1 0.1 0.1 0.1 0.01

1,2,3,4,7,8,9-HpCDF OCDF

81

92

ND ND ND ND ND 314 (61) 1742 (199) 89 (154) 36 (62)

– – – – – J(3),E(1) JB(3) J(1),ND(2) J(1),ND(2),E(1)

ND ND ND ND ND 387 (95) 2200 (428) 77 (134) ND

– – – – – J(3),E(1) JB(2),B(1) J(1),ND(2) –

ND ND ND ND ND 185 (46) 893 (200) 32 (55) 16 (27)

ND 21 (36) ND ND ND ND

– – – – – J(3),E(1) JB(3),E(1) JC(3) J(1),ND(2), E(1) – JB(1),ND(2),E(1) – – – –

32 (55) 44 (4) ND ND ND ND

J(1),ND(2) JB(3),E(1) – – – –

ND 11 (18) ND ND ND ND

– J(1),ND(2) – – – –

0.01 0.0003

ND ND

– –

ND 43 (75)

– J(1),ND(2)

ND 40 (69)

PCB81 PCB77 PCB123

0.00030 0.00010 0.00003

ND ND ND

– – –

ND ND ND

– – –

PCB118 PCB114

0.00003 0.00003

14389 (8741) ND

QB(2),B(1) –

1555 (2448) ND

PCB105

0.00003

1222 (2117)

JB(1),ND(2)

PCB126 PCB167 PCB156

0.10000 0.00003 0.00003

ND ND ND

PCB157 PCB169 PCB189 PCDD-TEQ PCDF-TEQ PCB-TEQ TOTAL TEQ

0.00003 0.03000 0.00003

ND ND ND 3.2 (0.4) 25.2 (5.0) 0.5 (0.2) 28.9 (5.4)

K1R5F ND ND ND 73 (64) 79 (68) 1502 (59) 6911 (598) 51 (88) ND

– – – J(2),ND(1) J(2),ND(1) – – ND(2),E(1) –

ND ND ND ND ND ND

– – – – – J(3),E(2) B(2),JB(1) JB(1),ND(2) J(1),ND(2), E(1) – – – – – –

70 (61) ND ND ND ND 34 (59)

ND ND

– –

ND 123 (214)

ND ND ND

– J(1),ND(2), E(1) – – –

J(2),ND(1) – – – – J(1),ND(2), E(1) – J(1),ND(2)

ND ND 211 (365)

ND 3610 (766) 21400 (3378)

– J(3),E(1) –

B(3) –

5500 (4869) ND

B(1),QB(2) –

8045 (389) ND

– – JB(1), ND(2) B(3),E(1) –

125000 (15308) 1767 (3060)

ND

–

2067 (1841)

1967 (185)

JB(2),QJB(1)

46444 (7706)

– – –

ND ND ND

– – –

ND 4000 (6928) 422 (732)

QJB(1),ND(1), JB(1),E(1) – QJB(1),ND(2) QJB(1),ND(2)

– J(1),ND(2), E(1) –

ND ND 811 (707)

ND 4056 (419) 7444 (193)

– QJB(3) QB(3)

– – –

ND ND ND 3.7 (0.6) 23.9 (16.6) 0.1 (0.1) 27.6 (16.7)

– – –

ND ND ND 4.5 (1.1) 8.8 (15.2) 0.4 (0.3) 13.7 (15.7)

– – –

ND ND ND 2.1 (0.5) 3.7 (5.2) 0.3 (0.1) 6.1 (5.4)

– – QJB(2), ND(1) – – –

ND ND ND 32.3 (13.3) 26.6 (24.2) 6.6 (0.7) 65.4 (37.6)

– – –

Results are given as mean (SD) of three samples for each brand on a fg/cigarette basis. Data flags are indicated and include the number of samples qualified by that flag in parenthesis. –: no data flag; ND: not detected; B: analyte detected in blank; C: value reported from confirmatory analysis; E: EMPC value; J: estimated value—below calibration range; Q: quantitative interference present. As discussed in the text, non-detects were assigned a value of zero for TEQ calculation purposes. Data qualified by other flags, including EMPCs, were used at the reported value. Congener specific values from a laboratory blank filter pad were subtracted from results from each sample.

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8-PeCDF

80

Table 3B Analytical results for Medium Yield cigarette brands Analyte

TEF

Cigarette Identification (Medium Yield) 78

85

86

1 1 0.1 0.1 0.1 0.01 0.0003 0.10 0.03 0.30 0.1 0.1 0.1 0.1 0.01

ND ND ND ND ND 211 (365) 4122 (3076) 154 (267) ND ND ND ND ND ND ND

– – – – – J(1),ND(2) JB(2),B(1) J(1),ND(2) – – – – – – –

ND ND ND ND ND 622 (47) 2744 (875) 611 (104) 218 (189) 230 (200) 178 (155) ND ND ND ND

– – – – – J(3) JB(2),B(1) JC(1),C(2),E(1) J(2),ND(1),E(1) J(2),ND(1),E(1) J(2),ND(1) – – – –

ND ND ND ND ND 813 (534) 6497 (4618) 161 (278) ND ND 92 (81) ND ND ND 1333 (1155)

1,2,3,4,7,8,9-HpCDF OCDF PCB81 PCB77 PCB123 PCB118

0.01 0.0003 0.00030 0.00010 0.00003 0.00003

ND ND ND ND ND 2778(4811)

– – – – – B(3)

ND 167 (289) ND ND ND 822 (719)

–

ND 274 (400) ND ND ND 356 (616)

PCB114 PCB105 PCB126 PCB167 PCB156 PCB157 PCB169 PCB189 PCDD-TEQ PCDF-TEQ PCB-TEQ TOTAL TEQ

0.00003 0.00003 0.10000 0.00003 0.00003 0.00003 0.03000 0.00003

ND 3545 (3160) ND ND ND ND ND ND 3.3 (4.6) 15.4 (26.7) 0.2 (0.2) 19.0 (24.6)

– JB(2),ND(1),E(1) – – – – – –

ND 3489 (4305) ND ND 1278 (2213) ND ND ND 7.0 (0.7) 154.6 (75.5) 0.17 (0.1) 161.8 (75.8)

– – – B(3),E(1) – JB(2),ND(1) – – QJB(1),ND(2) – – –

ND ND ND ND ND ND ND ND 10.1 (6.7) 26.7 (30.5) 0.0 (0.0) 36.8 (25.0)

87 – – – – – JB(3) JB(1),B(2) J(1),ND(2) – – J(2),ND(1) – – – J(2),ND(1), E(1) – JB(3) – – – B(2),JB(1), E(1) – – – – – – – –

K2R4F

ND ND ND ND ND 977 (631) 7533 (4612) 185 (320) ND ND ND ND ND ND 117 (203) ND 1213 (1315) ND ND ND 6189 (10006) ND ND ND ND ND ND ND ND 12.0 20.0 0.19 32.2

(7.7) (31.0) (0.3) (29.3)

– – – – – J(3) B(2),JB(1) J(1),ND(2) – – – – – – J(1),ND(2), E(1) – J(2),ND(1) – – – B(3)

ND ND ND ND ND 1039 (80) 6844 (767) ND ND ND ND ND ND ND ND

– – – – – J(3) B(3) – – – – – – – –

ND 423 (497) ND ND ND 9845 (9179)

– J(2),ND(1) – – – B(3),E(2)

– – – – – – – –

ND ND ND ND ND ND ND ND 12.4 (1.0) 0.1 (0.1) 0.3 (0.3) 12.9 (1.2)

– – – – – – – –

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF

Results are given as mean (SD) of three samples for each brand on a fg/cigarette basis. Data flags are indicated and include the number of samples qualified by that flag in parenthesis. –: no data flag; ND: not detected; B: analyte detected in blank; C: value reported from confirmatory analysis; E: EMPC value; J: estimated value—below calibration range; Q: quantitative interference present. As discussed in the text, non-detects were assigned a value of zero for TEQ calculation purposes. Data qualified by other flags, including EMPCs, were used at the reported value. Congener specific values from a laboratory blank filter pad were subtracted from results from each sample.

1727

1728

Table 3C Analytical results for High Yield cigarette brands Analyte

TEF

Cigarette Identification (High Yield) 88

1 1 0.1 0.1 0.1 0.01 0.0003 0.10 0.03 0.30 0.1 0.1 0.1 0.1 0.01 0.01 0.0003 0.00030 0.00010 0.00003 0.00003 0.00003 0.00003 0.10000 0.00003 0.00003 0.00003 0.03000 0.00003

ND ND ND ND ND 632 (433) 4322 (3074) 310 (268) 195 (173) 153 (140) 121 (105) ND ND ND 279 (254) ND 733 (1126) ND ND ND ND ND ND ND ND ND ND ND ND 7.7 (5.3) 97.9 (58.8) 0.0 (0.0) 105.5 (64.0)

– – – – – JB(3) JB(2),B(1) JC(2),ND(1) J(2),ND(1) J(2),ND(1),E(2) J(2),ND(1) – – – J(2),ND(1) – JB(3) – – – QJB(2),JB(1) E(2) – – – – – – – –

ND ND ND ND ND 421 (146) 4121 (1453) 683 (448) 171 (150) 179 (4) 198 (12) 34 (30) ND ND 243 (20) ND 170 (115) ND 3667 (6351) ND 14667 (12979) ND 6556 (11355) ND ND 7444 (12894) ND ND ND 5.4 (1.9) 152.7 (38.7) 1.2 (1.6) 159.4 (47.7)

79 – – – – – JB(3) JB(3) JC(2),E(2) J(2),ND(1) J(3),E(1) J(3) J(2),ND(1) – – J(3) – JB(3),E(1) – J(1),ND(2),E(1) – B(3),E(1) – JB(1),ND(2)E(1) – – QJ(1),ND(2) E(1) – – –

102 (176) ND ND 93 (162) ND 540 (64) 3633 (208) 451 (405) 187 (323) ND 33 (57) 8 (14) ND ND 183 (192) ND 373 (81) ND ND ND ND ND 2756 (4773) ND ND ND ND ND ND 117.5 (192.8) 64.0 (56.5) 0.1 (0.1) 181.7 (237.4)

89 J(1),ND(2),E(1) – – J(1),ND(2) – JB(3) JB(3) J(1),JC(1),ND(1) J(1),ND(2) – J(1),ND(2),E(1) J(1),ND(2) – – J(2),ND(1) – JB(3),E(1) – – – QJB(3) – QJB(1),ND(2),E(1) – – – – – –

ND ND ND ND ND 532 (174) 4247 (453) 493 (39) ND ND 40 (69) ND ND ND 114 (101) ND 17 (29) ND ND ND 15478 (10854) ND ND ND ND ND ND ND ND 6.6 (1.8) 54.4 (8.5) 0.5 (0.3) 61.5 (10.6)

– – – – – JB(3) JB(3) J(3) – – J(1),ND(2) – – – J(2),ND(1) E(1) – JB(3),E(1) – – – B(3),E(1) – – – – – – – –

Results are given as mean (SD) of three samples for each brand on a fg/cigarette basis. Data flags are indicated and include the number of samples qualified by that flag in parenthesis. –: no data flag; ND: not detected; B: analyte detected in blank; C: value reported from confirmatory analysis; E: EMPC value; J: estimated value—below calibration range; Q: quantitative interference present. As discussed in the text, non-detects were assigned a value of zero for TEQ calculation purposes. Data qualified by other flags, including EMPCs, were used at the reported value. Congener specific values from a laboratory blank filter pad were subtracted from results from each sample.

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF PCB81 PCB77 PCB123 PCB118 PCB114 PCB105 PCB126 PCB167 PCB156 PCB157 PCB169 PCB189 PCDD-TEQ PCDF-TEQ PCB-TEQ TOTAL TEQ

91

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

or exclusion of EMPC values in the calculation with the exception of a clear disparity between TEQ values calculated with and without EMPCs for Brands 85, 88, and 91. For Brand 85, inclusion of EMPCs in calculating TEQs for 2,3,7,8-TCDF, 2,3,4,7,8-PeCDF, and 1,2,3,7,8-PeCDF resulted in the higher observed TEQ concentration on both a per mg TPM basis (Fig. 2A) and per cigarette basis (Fig. 2B). As the 2005 WHO TEFs for these three congeners are 0.1, 0.3, and 0.03, respectively, these constituents carry substantial weight in the overall TEQ calculation. Similarly, using EMPCs for OCDD and 2,3,4,7,8-PeCDF in calculating TEQs for Brand 88 resulted in a higher observed TEQ concentration on both a per mg TPM basis (Fig. 2A) and per cigarette basis (Fig. 2B). The TEQ concentration on a per mg TPM and per cigarette basis was likewise increased when EMPCs for TCDF, 2,3,4,7,8-PeCDF, 1,2,3,4,7,8-HxCDF, OCDF, PCB77, PCB118, PCB105, and PCB156 were used in the calculations for Brand 91. To clarify further, in calculations where EMPCs were not used to estimate TEQs (shown by the non-shaded bars in Figs. 2 and 3), the concentration of a specific congener in at least one of the replicates would have taken on a value of zero and be eliminated from the calculation. As discussed above, this approach is consistent with EPA guidance regarding the use of EMPCs in calculation of TEQs (USEPA, 1996). The results in Fig. 3 show the relationship between the three ‘tar’ categories and TEQ concentration on a per mg TPM basis (Fig. 3A), a per cigarette basis (Fig. 3B), and per kg bw/day basis (Fig. 4). When normalized on a per mg TPM basis, no comparisons between ‘tar’ yield categories were statistically significant when calculated both with and without EMPC values, although the mean TEQ yield for the LY cigarette brands was significantly lower than for the 1R5F reference cigarette (p < 0.05). When compared on a per cigarette basis (Fig. 3B) and per kg bw/ day basis (Fig. 4), mean TEQ yields in the LY cigarette brands were significantly lower than for the HY cigarette brands, although mean TEQ yields for the LY brands were not lower than for the 1R5F low ‘tar’ yield reference cigarette. The statistical relationship between ‘tar’ level and TPM and fg TEQ/cig is reported in Table 5. Correlation analysis showed a significant positive correlation with both ‘tar’ and TPM for TEQ yield per cigarette and TEQ yield on a per kg body weight/day basis. As in the analyses described above, correlation analysis was also performed both with and without EMPCs. The outcome did not change: correlation analysis showed significant positive correlation regardless of whether EMPCs were included in the calculation. 4. Discussion The presence of PCDDs in cigarette MSS was described initially by Bumb et al. (1980), who reported the presence of the Hx-, Hp-, and OCDD congeners at trace concentra-

1729

tions (0.004–0.05 ppb). Subsequent studies by Muto and Takizawa reported the presence of predominately higher chlorinated (i.e., HxCDD, HpCDD, OCDD) 2,3,7,8PCDDs in both unburned cigarettes and in cigarette smoke (1989). Notably, neither of these studies utilized a standardized smoking regimen and neither analyzed for PCDFs or PCBs. However, several subsequent studies that have used standardized smoking protocols and have analyzed for both PCDDs and PCDFs report detecting HpCDD, OCDD, and HpCDF as the predominant congeners (Ball et al., 1990; Lo¨froth and Zebhr, 1992; Smith et al., 2004). PCDDs and PCDFs are formed from incomplete combustion processes of organic material, whether anthropogenic or natural, and are ubiquitous contaminants of the ecosystem, including commodity crops like tobacco. As such, the major source of dioxins in cigarette MSS likely arises from transfer of existing compounds from the tobacco to MSS, with evolution through pyrosynthesis as a minor pathway. In addition to the reports describing the presence of dioxins in MSS, PCDDs, PCDFs, and PCBs have been identified in unburned tobacco (Aoyama et al., 2003; Matsueda et al., 1994; Smith et al., 2004; Muto and Takizawa, 1989). Aoyama et al. (2003) attempted a mass balance estimate by subtracting the sum of the dioxin concentrations in the filter after smoking and in the ash from the concentrations determined in the tobacco leaf and paper. Using their mass balance approach, but no empirically derived data, approximately 63% of dioxins in the unburned cigarette can be found in the MSS or sidestream smoke. Similarly, Matsueda et al. (1994) concluded that PCDDs/PCDFs in tobacco transfer to MSS at a rate of about 35%. Interestingly, the concentration of dioxins in unburned cigarettes did not vary considerably across a range of ‘tar’ deliveries, and all dioxins were attributed to tobacco filler, not cigarette paper (Smith et al., 2004). ‘Tar’ yield in a given cigarette is influenced not only by tobacco weight and type, but by the machine smoking regimen employed and a multitude of design factors, including paper porosity, ventilation, pressure drop, and filter length. As such, it is reasonable to conclude that the dioxin concentrations in unburned tobacco are relatively uniform across ‘tar’ categories and this is consistent with our data showing no differences between ‘tar’ categories when dioxin yield is normalized to TPM (Figs. 2A and 3A). In contrast, but consistent with the hypothesis that other design features influence ‘tar’ yield, MSS dioxin concentrations expressed on a per cigarette basis do increase with ‘tar’ yield. Congener patterns reported in studies of dioxins in unburned cigarettes vary widely by brand (Aoyama et al., 2003). For example, in a Japanese brand, concentrations of PCDDs in tobacco leaf and paper were the highest among congeners, while American and British cigarettes tested in the same study contained OCDD as the major congener (Aoyama et al., 2003). This variation was also observed by Matsueda et al. (1994), who detected OCDF

1730

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

Table 4 Summary of results from eight studies of dioxin concentrations in cigarettes or cigarette smoke Study

Smoking regimen

Analytes

Results Qualitative

Quantitative (mean pg TEQ/cig)

Bumb et al. (1980)

20–30 puffs, 2–3 s in duration

PCDD

0.895 pg TEQ/g smoke

Muto and Takizawa (1989)

Continuous smoking

PCDD

Ball et al. (1990) Lo¨froth and Zebhr (1992) Matsueda et al. (1994)

35/58/2 (DIN 10 240) 35/60/2 (ISO) Cigarettes not smoked

PCDD/PCDF PCDD/PCDF PCDD/F/PCB

Aoyama et al. (2003) Smith et al. (2004)

Cigarettes not smoked 35/60/2 (ISO)

PCDD/F/PCB PCDD/F/PCB

Present study

60/30/2

PCDD/F/PCB

HxCDD HpCDD OCDD HxCDD HpCDD OCDD HpCDD OCDD HpCDD OCDD HpCDF PCB77 OCDD HpCDF OCDF OCDD HpCDD OCDD HxCDF PCB105 HpCDD OCDD OCDF PCB105 PCB118

3.7

0.06 0.9 0.4

1.3 0.01 (ULT) 0.01 (LT) 0.03 (FF)

0.02 (LY) 0.05 (MY) 0.08 (HY)

‘‘Qualitative” data includes information on the predominant congeners (2,3,7,8-substituted PCDDs/PCDFs or non- or mono-ortho-substituted PCBs) from a particular study. ‘‘Quantitative” data are expressed as mean pg TEQ/cigarette and represents the average concentration calculated based on reported values in each study. Aoyama et al. (2003) estimated smoke concentrations through mass-balance calculations, but did not measure directly. Smoking regimen values are designated as puff volume (ml)/puff frequency (s)/puff duration (s). To allow for consistency with comparison to other studies, results from the present study were calculated using only values that exceeded both the LOD and LOQ (i.e., did not include EMPCs).

as the major PCDF analyte in unburned cigarettes from the United States, Taiwan, and Denmark, while cigarettes from the United Kingdom, Germany, Japan, and China exhibited broad variability in concentration of the 2,3,7,8-substituted PCDF congeners (Matsueda et al., 1994). These qualitative analyses are summarized in Table 4 and are consistent with the congener pattern observed in MSS in the present study. Fig. 1 shows the relative contribution to TEQs on a per cigarette basis from PCDDs, PCDFs, and PCBs. PCDF congeners contribute the greatest proportion to the total TEQs. Interestingly, PCB concentrations decrease with increasing ‘tar’ yield. The 1R5F and 2R4F Kentucky reference cigarettes, LY and MY reference cigarettes, respectively, also follow this general trend: as ‘tar’ yield increases, TEQs attributable to PCB congeners decrease. Smith et al. (2004) published the first report exploring the relationship between TPM yield and dioxin concentrations in cigarette MSS. Consistent with the study by Smith, results in Fig. 3 show that MSS TEQ concentration on a per cigarette basis is not significantly different between LY and MY cigarettes, but that HY cigarettes do deliver a significantly higher concentration of dioxins (vs. LY). In contrast to their results, we found a significant, positive correlation between both TPM and ‘tar’ yield and dioxin concentration (Table 5). Likewise, we found no statistically significant relationship between ‘tar’ yield and TEQ con-

centration when adjusted per mg TPM (Figs. 2 and 3). Finally, we report lower TEQ concentrations for all ‘tar’ yields when adjusted per mg TPM. There are several potential reasons why some of the results and conclusions presented herein do not agree with those of Smith et al. (2004). First, TEFs were recently reevaluated by a WHO International Programme on Chemical Safety expert panel (Van den Berg et al., 2006), resulting in differences in calculated TEQs. Second, none of the cigarettes in the present study contained charcoal filters, whereas all of the cigarettes tested by Smith had charcoal filters. The use of a charcoal filter could possibly reduce the transfer of PCBs from the tobacco to MSS and subsequently in TPM, resulting in lower TEQs. Third, Smith and colleagues report that all cigarettes were ‘‘American-style blend”, but only three were actually manufactured in the United States, while the country of origin of the other four brands was not disclosed. There are significant quantitative and qualitative differences among products from cigarette manufacturers around the world, and differences in tobacco type and source, structural materials, functional design, physical dimensions, filter type, tobacco blend, moisture content, and added flavorings can and do have significant impacts on MSS composition and yield (Borgerding and Klus, 2005). In the present study, 12 brands from the same US manufacturer were used.

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733 Table 5 Analysis of correlation between FTC ‘tar’, TPM, and measures of TEQ concentration Without EMPC

With EMPC

r

p: slope = 0

r

p: slope = 0

‘Tar’ vs. fg TEQ/mg TPM fg TEQ/cig pg TEQ/kg bw/day

0.424 0.802 0.802

0.17 0.0017 0.0017

0.303 0.732 0.732

0.34 0.0068 0.0068

TPM (60/30/2) vs. fg TEQ/mg TPM fg TEQ/cig pg TEQ/kg bw/day

0.451 0.822 0.822

0.14 0.0010 0.0010

0.402 0.779 0.779

0.20 0.0028 0.0028

Correlation coefficient (r) and statistical significance were calculated both with and without Estimated Maximum Possible Concentrations (EMPCs). Values were log-transformed to make within-cigarette variation more consistent among cigarettes. Means in log units were then back-transformed prior to correlation analysis. Correlation analysis still showed significant, positive correlation when calculated using ‘‘as observed” values vs. log-transformed data, although correlation coefficients were lower (data not shown).

Finally, the present study used an ‘‘intense” smoking regimen (60/30/2), whereas Smith et al. (2004) used the less intense ISO smoking protocol (35/60/2). While neither regimen exactly replicates the dynamic nature of smoking in humans, the more intense smoking regimen produces ‘tar’ yields that are greater than average daily yields experienced by 95% of smokers when smoking their usual MY and HY cigarette brands (Borgerding and Klus, 2005). Likewise, the intense regimen produces greater ‘tar’ yields than average daily yields experienced by approximately 60% of smokers when smoking their usual LY brand. In contrast, the ISO regimen produced ‘tar’ yields that were greater than average daily yields experienced by only 4% of smokers smoking LY, around 15% of subjects smoking their usual MY cigarette brands, and less than 50% of smokers smoking their usual HY cigarettes (Borgerding and Klus, 2005). Importantly, the LY and MY categories constituted six of the seven brands studied by Smith et al. (2004). While no machine smoking protocol can mimic the interindividual variability in human smoking behavior, the ISO regimen used by Smith and colleagues likely underestimates average exposure, while the intense regimen used in the present study likely overestimates typical exposure per cigarette. With further regard to differences in the two smoking regimens, more intensive smoking may result in more complete combustion, but whether dioxin concentrations change solely as a function of regimen cannot be concluded in the present study (Dixon and Borgerding, 2006). TEQ yields for the 1R5F are nearly 12-fold greater on a per mg TPM basis and over 3-fold greater than those of 2R4F on a per cigarette basis (p < 0.05) (Figs. 3A and 3B). The 1R5F Kentucky reference cigarette was developed in 1989 by the University of Kentucky Tobacco and Health Research Institute as a LY reference cigarette for experimental research purposes, yielding approximately 1.7 mg FTC ‘tar’ per cigarette (Diana and Vaught, 1990). The 2R4F Kentucky reference cigarette was developed and

1731

manufactured in 2005 as a MY reference cigarette and yields nominally 9 mg FTC ‘tar’. Collectively, these US blended reference cigarettes represent some of the major design features introduced in the last two decades: novel filter design and ventilation, use of tobacco ‘‘sheet”, and, in the case of 1R5F, use of expanded tobacco. Clearly, the TEQ yields for these cigarettes differ substantially even when normalized for TPM yield (Fig. 3A), which is not observed for the three test categories. The 2R4F was produced over fifteen years later than the 1R5F. During this time, there was an apparent decline in environmental dioxin burden and consequent reduced uptake or deposition on tobacco plants. This evolution may partially explain the differences in TEQ yields. USEPA (2000) reported that TCDD emissions from quantified sources decreased by over 4-fold between 1987 and 1995. This has been accompanied by a decrease in PCDD/PCDF body burdens (Aylward and Hays, 2002) and estimated intake (Charnley and Doull, 2005). Potential sources of emissions during the period between 1950 and the mid-1980s include the use of herbicides with trace dioxin contamination, combustion of leaded gasoline, uncontrolled open burning of landfills and refuse, municipal waste incinerators, agricultural burning practices, application of animal waste products to soil as fertilizers, and forest and brush fires (reviewed in Hays and Aylward, 2003), all of which could have taken place in conjunction with or in the vicinity of tobacco agricultural practices. The data presented in Table 4 from previous studies of dioxins in cigarettes or cigarette smoke is largely insufficient to be able to draw conclusions specifically about historical trends. For example, the available studies that reported dioxin concentrations in unburned tobacco (Matsueda et al., 1994; Aoyama et al., 2003; Smith et al., 2004) do not support the hypothesis that levels in unburned tobacco have decreased since the mid-1990s. However, these studies are difficult to compare given that the application of TEFs to analytical data differed, as did the origin of cigarettes, collection methods, and cigarette design criteria (i.e., full-flavor, etc.) were only reported in one of the studies (Smith et al., 2004). Comparing historical data from tobacco smoke is equally difficult to interpret for many of the same reasons: cigarettes of different origin, different design characteristics, inconsistent application of TEFs to analytical data and, most importantly, inconsistent use of machine smoking regimens. The two studies that did use the same smoking regimen (Lo¨froth and Zebhr, 1992; Smith et al., 2004) of 35/60/2 (ISO) differed in publication date by 12 years and did show an approximate 30-fold decrease in mean pg TEQ/cigarette for full-flavor cigarettes from 1992 to 2004. Comparing the results of the present study to the data reported by Lo¨froth and Zebhr (1992) also shows a decrease in pg TEQ/cig for high-tar (HY, full-flavor) cigarettes since 1992. However, because a more intense smoking regimen was used in the present study, the data cannot be compared as readily as with that reported by Smith et al. (2004).

1732

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733

The use of the TEF approach in risk assessment of abiotic matrices or when using routes of exposure other than the oral route is controversial (Van den Berg et al., 2006), adding uncertainty to risk assessment using the present data set. The compounds of interest in this study are highly hydrophobic and strongly associated with particulate matter. As such, bioavailability of dioxins from cigarette smoke in the lung is governed by several physicochemical thresholds: retention of inhaled particles from cigarette smoke by the lung, subsequent dissociation of dioxins from retained particulate matter, and alveolar uptake of dissociated dioxins. Addressing any of the three in great detail is beyond the scope of this study; however, the particulate phase of cigarette MSS contains particles of diverse composition and size range. Physical characteristics of particles that influence rate, site, and extent of retention include particle shape, physical size, and density, along with lung anatomy and airflow patterns, determine the fate of inhaled particles in the respiratory tract (Bogdanffy and Keller, 1999). Lipophilic compounds are rapidly absorbed from the alveolar spaces, and the dioxins are highly bioavailable even when bound to respirable soil particles (Nessel et al., 1992). Despite the uncertainties associated with use of TEFs in risk assessment of cigarette smoke, the use of the TEQ is a useful and accepted tool for comparing potential exposure between similar matrices (Van den Berg et al., 2006). The major uncertainties associated with risk assessment of dioxins in cigarette MSS, in addition to the inherent uncertainties associated with the use of TEFs, include the fact that cigarette MSS is a highly complex mixture. In this study, only compounds for which a TEF had been assigned Van den Berg et al. (2006) were analyzed. However, other unmeasured components may also contribute to the overall biological response to aromatic hydrocarbons, including numerous PAHs. Importantly, several naturally occurring compounds, such as indole-3-carbinol (I3C) and related hetero-PAHs, are known to bind the Ah receptor, potentially mitigating AhR-mediated biochemical and toxic effects induced by dioxins (Safe, 1995; Safe, 1998). The dioxin exposure estimates for MSS calculated using the TEF approach are consistent with previous observations in that these compounds have been identified in unburned tobacco and mainstream smoke at relatively low levels (Matsueda et al., 1994; Aoyama et al., 2003; Lo¨froth and Zebhr, 1992; Smith et al., 2004; Muto and Takizawa, 1989; Ball et al., 1990; Bumb et al., 1980) (Table 4). In the present study, daily exposure estimates fall several orders of magnitude below the current WHO tolerable daily intake (TDI) range of 1–4 pg/kg-bw/day (van Leeuwen et al., 2000) (Fig. 4) even when calculated using 38 cigarettes per day (95th percentile of cigarettes smoked per day as measured in the 1999–2000 NHANES survey) and a body weight of 75 kg (median as measured in the 1999–2000 NHANES survey) (National Center for Health Statistics, 1999). US EPA estimates of mean dietary intake for the mid-1990s of PCDDs/PCDFs and dioxin-like PCBs

are 0.6 pg TEQ/kg bw/day and 0.34 pg/kg bw/day, respectively, for a total of about 1 pg TEQ/kg bw/day (USEPA, 2000). Beef, milk, and dairy account for nearly 50% of the TEQ intake, while vegetables and grains are generally low in dioxin content (USEPA, 2000). The dioxins from sources other than food are not calculated in these ‘‘market basket” studies. However, given recent estimates of total dioxin intake from food, cigarette smoking, even at very high levels of exposure, provides a minimal, although measurable, contribution to exposure to dioxins. Conflict of interest statement All work was funded internally by R.J. Reynolds Tobacco Company. Authors report no conflicts of interest. Acknowledgement The authors wish to acknowledge the excellent work of the R.J. Reynolds Tobacco Company Analytical Laboratories and Eno River Laboratories Inc. We also thank our colleagues, Dr. Serban Moldoveanu and Dr. Suzana Theophilus, for their insightful and thorough reviews of the manuscript. References Aoyama, T., Ikeda, K., Takatori, A., Obal, T., 2003. Risk assessment of dioxins in cigarette smoke. Organohalogen Compd. 65, 321–324. Aylward, L.L., Hays, S.M., 2002. Temporal trends in human TCDD body burden: decreases over three decades and implications for exposure levels. J. Exp. Anal. Environ. Epidemiol. 12, 319–328. Ball, M., Pa¨pke, O., Lis, A., 1990. Polychlorodibenzodioxins and polychlorodibenzofurans in cigarette smoke. Bietr. Tabakforsch. Int. 14, 393–402. Ballschmiter, K., Rappe, C., Buser, H.R., 1980. Chemical properties, analytical and environmental levels of PCBs, PCTs, PCNs, and PCBs. In: Kimbrough, R.D., Jensen, A.A. (Eds.), Halogenated Biphenyls, Terphenyls, Napthalenes, Dibenzodioxins, and Related Products. Elsevier, Amsterdam, pp. 47–69. Bogdanffy, M.S., Keller, D.A., 1999. Metabolism of xenobiotics by the respiratory tract. In: Gardner, D.E., Crapo, J.D., McClellan, R.O. (Eds.), Toxicology of the Lung, third ed. Taylor & Francis, Philadelphia, pp. 85–124. Borgerding, M., Klus, H., 2005. Analysis of complex mixtures – cigarette smoke. Exp. Toxicol. Pathol. 57, S43–S73. Brinkman, U.A.T., DeKok, A., 1989. Production, properties, and uses. In: Kimbrough, R.D., Jensen, A.A. (Eds.), Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins, and Related Products. Elsevier, Amsterdam, pp. 1–40. Bumb, R.R., Crummet, W.B., Cutie, S.S., Gledhill, J.R., Hummel, R.H., Kagel, R.O., Lamparski, L.L., Luoma, E.V., Miller, D.L., Nestrick, T.J., Shadoff, L.A., Stehl, R.H., Woods, J.S., 1980. Trace chemistries of fire: a source of chlorinated dioxins. Science 210, 385–390. Charnley, G., Doull, J., 2005. Human exposure to dioxins from food, 1999–2002. Food Chem. Toxicol. 43, 671–679. Chepiga, T.A., Morton, M.J., Murphy, P.A., Avalos, J.T., Bombick, B.R., Doolittle, D.J., Borgerding, M.F., Swauger, J.E., 2000. A comparison of the mainstream smoke chemistry and mutagenicity of a representative sample of the US cigarette market with two Kentucky reference cigarettes (K1R4F and K1R5F). Food Chem. Toxicol. 28, 949–962. DeVoogt, P., Brinkman, U.A.T., 1989. Production, properties, and usage of polychlorinated biphenyls. In: Kimbrough, R.D., Jensen, A.A. (Eds.),

C.L. Wilson et al. / Food and Chemical Toxicology 46 (2008) 1721–1733 Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins, and Related Products. Elsevier, Amsterdam, pp. 3–45. Diana, J.N., Vaught, A., 1990. Research Cigarettes. University of Kentucky, Lexington, KY. Dixon, M., Borgerding, M.F., 2006. Recent advances in the application and understanding of alternative smoking regimes. Recent Advances in Tobacco Science 32, 3–83. Dube, M.F., Green, C.R., 1982. Methods of collection of smoke for research purposes. Recent Adv. Tobacco Sci. 8, 42–102. Hays, S.M., Aylward, L.L., 2003. Dioxin risks in perspective: past, present, and future. Regul. Toxicol. Pharmacol. 37, 202–217. Hoffmann, D., Hoffmann, I., 1997. The changing cigarette, 1950–1995. J. Toxicol. Environ. Health 50, 307–364. Kiviranta, H., Ovaskainen, M-L., Vartiainen, T., 2004. Market basket study on dietary intake of PCDD/Fs, PCBs, and PBDEs in Finland. Environ. Int. 20, 923–932. Lo¨froth, G., Zebhr, Y., 1992. Polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in mainstream and sidestream cigarette smoke. Bull. Environ. Contam. Toxicol. 48, 789–794. Matsueda, T., Kurokawa, Y., Nakamura, M., Takada, S., Fukamachi, K., 1994. Concentrations of PCDDs, PCDFs, and coplanar PCBs in cigarettes from various countries. Organohalogen Compd. 20, 331–334. Muto, H., Takizawa, Y., 1989. Dioxins in cigarette smoke. Arch. Environ. Health 44, 171–174. National Center for Health Statistics, Centers for Disease Control and Prevention. National Health and Nutrition Examination Survey. NHANES 1999–2000. Hyattsville, MD: National Center for Health Statistics. (http://www.cdc.gov/nchs/about/major/nhanes/nhanes99_00.htm). Accessed 7 March 2007. Nessel, C.S., Amoruso, M.A., Umbreit, T.H., Meeker, R.J., Gallo, M.A., 1992. Pulmonary bioavailability and fine particle enrichment of 2,3,7,8-tetrachlorodibenzo-p-dioxin in respirable soil particles. Fundam. Appl. Toxicol. 19, 279–285. Norman, A., 1999. Cigarette design and materials. In: Davis, D.L., Nelson, M.T. (Eds.), Tobacco: Production, Chemistry, and Technology. Blackwell Science, London, pp. 353–387. Pillsbury, H.C., Bright, C.C., O’Conner, K.J., Irish, F.W., 1969. Tar and nicotine in cigarette smoke. J. Assoc. Off. Anal. Chem. 52, 458–462. Rappe, C., Buser, H.R., Bosshardt, H.P., 1979. Dioxins, dibenzofurans and other polyhalogenated aromatics: production, use, formation, and destruction. Ann. NY Acad. Sci. 320, 1–18.

1733

Rodgman, A., 2001. Studies of polycyclic aromatic hydrocarbons in cigarette mainstream smoke: identification, tobacco precursors, control of levels: a review. Beitr. Tabakforch. Int. 19, 361–379. Rodgman, A., Green, C.R., 2003. Toxic chemicals in cigarette mainstream smoke – hazard and hoopla. Beitr. Tabakforsch. Int. 20, 481–545. Safe, S., 1992. Development, validation, and limitations of toxic equivalency factors. Chemosphere 25, 61–64. Safe, S., 1995. Development, validation, and problems with the Toxic Equivalency Factor approach for risk assessment of dioxins and related compounds. J. Anim. Sci. 76, 134–141. Safe, S.H., 1998. Hazard and risk assessment of chemical mixtures using the toxic equivalency. Environ. Health Perspect. 106 (Suppl 4), 1051– 1058. Schecter, A., Birnbaum, L., Ryan, J.J., Constable, J.D., 2006. Dioxins: an overview. Environ. Res. 101, 419–428. Smith, C.J., Sykes, D.C., Cantrell, D.W., Moldoveanu, S., 2004. Dioxin levels in mainstream smoke from cigarettes with different TPM deliveries. Bietr. Tabakforsch. Int. 21, 205–209. US EPA, 1996. Test Methods for Evaluating Solid Waste, Physical/ Chemical Methods. Office of Solid Waste. US Environmental Protection Agency, Washington, D.C. US EPA, 2000. Exposure and human health reassessment of 2,3,7,8tetrachlorodibenzo-p-dioxin and related compounds. National Center for Environmental Assessment, US Environmental Protection Agency, Washington, DC. US Food and Drug Administration (USFDA), 2004. PCDD/PCDF Exposure Estimates. Center for Food Safety and Applied Nutrition, Office of Plant and Dairy Foods. Washington, DC: US Food and Drug Administration. (http://www.cfsan.fda.gov/ ~ lrd/dioxee.html). Accessed 15 October 2007. Van den Berg, M., Birnbaum, L., Denison, M., De Vito, M., Farland, W., Feeley, M., Fiedler, H., Hakansson, H., Hanberg, A., Haws, L., Rose, M., Safe, S., Schrenk, D., Tohyama, C., Tritscher, A., Tuomisto, J., Tysklind, M., Walker, N., Peterson, R.E., 2006. The 2005 World Health Organization reevaluation of human and mammalian Toxic Equivalency Factors (TEFs) for dioxins and dioxin-like compounds. Toxicol. Sci. 93, 223–241. van Leeuwen, F.X., Feeley, M., Schrenk, D., Larsen, J.C., Farland, W., Younes, M., 2000. Dioxins: WHO’s tolerable daily intake (TDI) revisited. Chemosphere 40, 1095–1101.