WO2012110819A1 - Smoking articles - Google Patents

Abstract

A smoking article comprising at least two of: (a) a tobacco blend comprising one or more tobaccos or tobacco grades with low TSNA and/or metal content; (b) a tobacco blend that has been treated to remove polyphenols and/or peptides; (c) a tobacco substitute sheet comprising a non-combustible inorganic filler, a binder and an aerosol generating means; (d) a high activity carbon comprising a polymer-derived carbon material; and (e) an amine-functionalised chelating resin.

Description

Smoking Articles

Technical Field

The present invention relates to smoking articles and, m particular, to smoking articles which combine two or more technological applications that individually reduce the machine measured yields of specific constituents or groups of

constituents in mainstream smoke.

Background

Tobacco smoke is a complex, dynamic mixture of more than 5000 identified constituents of which approximately 150 have been documented as being

undesirable. The constituents are present in the mainstream smoke (MS) which is inhaled by a smoker and are also released between puffs as constituents of sidestream smoke (SS).

In 2001 the Institute of Medicine (IOM) reported that, since smoking related diseases were dose-related, and because epidemiologic studies show reduction m the risk of smoking related diseases following cessation, it might be possible to reduce smoking related risks by developing potential reduced-exposure products (PREPs). These they defined as: (1) products that result in the substantial reduction in exposure to one or more tobacco toxicants; and (2) if a risk reduction claim is made, products that can reasonably be expected to reduce the risk of one or more specific diseases or other adverse health effects (Stratton et al, 2001). To date, no

combustible cigarette product has been shown to meet the general requirements outlined by the IOM.

The IOM and other groups (Life Sciences Research Office (LSRO) 2007; World Health Organisation (WHO) 2007) describe a number of stages of activity which are likely to be required for a combustible tobacco product to be recognised as a PREP; however, the detailed approach and stages required to provide relevant data have yet to be agreed amongst the scientific community. For example, some groups have proposed MS yield limits for specific smoke constituents and others have suggested that biomonitoring should play a role m this assessment Much research has been done into the reduction of specific MS constituents over recent years. Approaches have taigeted different parts of the smoking article. There have been efforts to reduce the levels of or to remove certain compounds from the starting material, for example by genetic engineering or by blending of specific tobaccos. Tobacco treatments have sought to reduce or remove compounds from tobacco material prior to incorporation into the smoking article. Various ways of "diluting" the tobacco in the tobacco rod of a smoking article have been attempted, using various types of diluents or fillers. Other approaches have involved ventilation of the smoking article, where ambient air is drawn into the smoking article to dilute the MS. Filtration is obviously another area where much work has been done to enhance the removal of MS constituents as they pass through the filter section of the smoking article. All of these individual measures have benefits, but they generally only address a small part of the picture.

A further issue to be addressed is the importance of pioducing a product which is acceptable to the consumer. Much of the sensory impact of a conventional smoking article is based upon the constituents of the MS. Removing some of these has the potential to provide the smoker with an unsatisfactory smoking experience.

There is, therefore, a challenge to provide a smoking article which shows significant reduction m emissions of all MS constituents considered to be undesirable.

However, individual measures to reduce certain constituents will frequently give rise to no reduction in other constituents and, in some cases, even an increase in the levels of others.

Oveiall reductions in smoking machine measured toxicant yields can be achieved by diluting the smoke using filter ventilation or using cigarette papers with high permeability, and, m the case of toxicants that are associated with the particulate phase of smoke, by increasing the filtration efficiency of the filter. For many years, governments and public health authoiities in various parts of the world considered lower ISO tar yielding cigarettes as a way to reduce the health risks of smoking for those smokers who do not quit smoking. However, this product modification approach has more recently been highly criticised. The Study Group on Tobacco Product Regulation (TobReg) of the World Health Organization has recently proposed a regulatory approach that would limit the yields of a selected group of specific smoke constituents. This group also recommended that the yields of constituents should be limited on the basis of their yields measured with an intense smoking machine regime and determined per mg of nicotine.

Appioaches to selectively reducing specific smoke constituents relative to machine measured tar and nicotine yields are very dependent upon the physiochemical nature of the individual constituents. Conventional cigarette design parameters offer limited scope for relative reductions in the smoke constituents. For example, by increasing the filter efficiency of a conventional cellulose acetate (CA) filter, the particulate phase constituents are reduced with the tar and nicotine and little or no selective reduction occurs. And, since cellulose acetate filters have little or no effect on volatile constituents, increasing filtration efficiency increases the ratios of their yields relative to tar and nicotine.

Increasing filter ventilation has varied effects on the smoke constituents. The absolute yields of all the smoke constituents are reduced, but, relative to tar or nicotine, yields of most of the particulate phase constituents are unchanged or may even be increased. The yields of some of the volatile constituents, such as ammonia and carbon monoxide, are reduced relative to both tar and nicotine, while the relative yields of some of the semivolatile constituents such as phenols are increased.

Many of the volatile vapoui phase components, such as the volatile aldehydes and hydrogen cyanide may be selectively reduced using adsorbent materials m the filter such as activated charcoal or certain resins. However, permanent gases, such as carbon monoxide and nitric oxide, are not amenable to adsorption at room temperature, and toxicants m the particulate phase cannot be selectively reduced by filtration since they are largely bound into the aerosol particles.

Since the 1950s, attempts have been made to selectively remove or reduce constituents from cigarette smoke. Adsorption by porous adsorbents is a possible means of removing some of the volatile constituents from smoke. Active Carbon (AC) is a nonselective adsorbent which is widely used in cigarette filters and can reduce a broad range of volatile smoke constituents to a significant extent via physisorption. However, the difficulty of this challenge should not be

underestimated. With cigarette smoke adsorbents there is a need to operate under high flow rate conditions (approximately 1 L per mm for typical machine-smokmg conditions), and therefore very short contact times between smoke constituent and filtei adsorbent (of the order of milliseconds). Adsorbents also need to function at the gas-sohd interface (i.e. not m solution) and m the presence of thousands of other chemicals in both vapour and particulate phases. Adsorbent surfaces are also susceptible to blocking by condensing smoke aerosol particles. For permanent gases, and smoke constituents with high vapour pressures at ambient temperatures such as formaldehyde, acetaldehyde or HCN, physical adsorption has been found to be less effective and alternative routes are required.

Cigaiette smoke contains a number of volatile aldehydes, both saturated compounds such as formaldehyde, acetaldehyde, propionaldehyde and butyraldehyde, and unsaturated compounds such as acrolein and crotonaldehyde. Carbonyls m cigarette smoke are mainly generated by combustion of a number of tobacco constituents, mostly carbohydrates. In particular it is thought that sugars are major sources of formaldehyde in cigarette smoke. Cellulose has been suggested to be the major precursor of mainstream smoke acetaldehyde. There are some data suggesting that glyceiol, a material sometimes added to tobacco as a humectant, is an additional precursor for acrolein. Although the boiling point of formaldehyde is sub-ambient, 30% of formaldehyde in the mainstream smoke exiting a filtered cigarette resides m the particulate phase and thus is not available for selective filtration at room temperature. Due to the presence of water vapour, formaldehyde m the particulate phase of smoke exists as the hydrated foim, CH₂(OH)₂. Acetaldehyde, one of the highest yield constituents of cigarette smoke, exists at or around its boiling point at ambient temperatures, and therefore has a very high vapour pressure. The combination of these two factors makes substantial removal of acetaldehyde from a smoke stream by filter additives a major challenge.

A promising approach to achieving substantial specific reductions in particulate constituents from a conventionally structured cigarette is to modify the tobacco. Substitution of different tobacco varieties into the blend can have an impact on yields of several smoke constituents. For example there are higher yields of the nitrogen containing smoke constituents from burley tobacco than from flue cured or oriental, and higher yields of formaldehyde and catechol from flue-cured tobaccos. However, decreases in one constituents or set of constituents are often offset by increases m other constituents. To avoid this it would be useful to be able to identify and remove precursors to smoke constituents from the tobacco leaf.

With the exception of the metallic constituents (chromium, nickel, arsenic, selenium, cadmium, mercury and lead) and some of the tobacco specific

nitrosamines (TSNAs), such as NAT and NAB, which are transferred directly from the leaf, the majority of the smoke constituents are formed by pyrosyn thesis from the leaf components. Thus, the major precursors for the volatile carbonyls, ben2o(a)pyrene, carbon monoxide, benzene and toluene are the structural carbohydrates such as pectin and cellulose as well as the sugars. The nitrogenous smoke constituents are formed from nitrogenous precursors in the leaf, and there is considerable evidence that protein and amino acid combustion contributes to the generation of several nitrogen containing smoke constituents on the Health Canada hst. Proteins and amino acids have been reported to be precursors for hydrogen cyanide, pyridine and qumoline, 2-ammonaphthalene and 4-aminobiphenyl. Tobacco protein is also strongly correlated with the formation of mutagenic heterocyclic amines and the resulting mutagenicity of smoke condensate m the TA98 Ames assay.

The polyphenols m tobacco are major precursors for phenolic smoke compounds. Chlorogenic acid, the most abundant polyphenol in flue-cured tobacco, is a major precursor for phenol, catechol and the substituted catechols, while hydroquinone has also been reported as a chlorogenic acid pyrolysis product. Rutin and caffeic acid also generate catechol and substituted catechols on pyrolysis but because of their low concentrations in tobacco and because of their lower pyrolytic yields their contributions to catechol in flue-cured tobacco smoke are much less than

chlorogenic acid. Resorcinol is known to be a major product from pyrolysis of rutin.

Detailed Description

The present invention provides combinations of bespoke tobacco blends with bespoke adsorbent filter additives, which result m a smoking article having a significant reduction in mainstream smoke constituents considered to be undesirable.

More specifically, the present invention provides a smoking article comprising at least two of:

(a) a tobacco blend comprising one or more tobaccos or tobacco grades with low TSNA and/or metal content;

(b) a tobacco blend that has been treated to remove polyphenols and/ or peptides (soluble and insoluble);

(c) a tobacco substitute sheet comprising a non-combustible inorganic filler, a binder and an aerosol generating means;

(d) a high activity carbon comprising a polymer-derived carbon material; and

(e) an amine-functionahsed chelating resin.

In a preferred embodiment, the smoking articles according to the invention have a reduction in at least 75%, preferably at least 90% and more preferably m all of the key constituents of mainstream smoke, as defined herein.

The so-called "key constituents" of MS referred to m connection with the present invention are those smoke constituents which have been identified m the literature as being undesirable (see, for example, The Scientific Basis of Tobacco Product Regulation: Report of a WHO Study Group (2007) WHO Technical Report Series 945, Geneva) and/or those whose yields have been analysed m the data provided herein (see, for example, Tables 6, 7 and 8).

The reduction is preferably determined using one of the smoking machine conditions set out m Table 3. Preferably, the reduced yields are measured under Health Canada Intense smoking machine conditions. The reduction m yield of the key constituents is preferably at least 5% oi at least 10% or more.

Preferably, where the smoking articles of the present invention include a tobacco blend comprising one or more tobaccos or tobacco grades with low TSNA and/ or metal content, they further comprise two or more other technologies listed as (b) to (e).

Brief Description of Figures

In order that the invention may be more fully understood, aspects and

embodiments thereof will be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:

Figure 1 shows Table 2, setting out the cigarette construction details.

Figure 2 shows Table 4, setting out the major constituent yields of test cigarettes using different smoking machine condition.

Figure 3 shows Table 5, setting out the blend metal and tobacco-specific nitrosamine contents.

Figure 4 shows Table 6, setting out the MS yields of metals and TSNAs measured under Health Canada Intense smoking machine conditions. Figures 5A, 5B and 5C show Table 7, setting out the MS yields of other smoke constituents measured under Health Canada Intense smoking machine conditions.

Figure 6 shows Table 8, setting out the MS yields of carbonyl and miscellaneous volatile and vapour phase smoke constituents in control and triple stage filter EC measured under Health Canada Intense smoking machine conditions. Figure 7 shows Table 9, setting out the MS yields of carbonyl and miscellaneous volatile and vapour phase smoke constituents in control and dual stage filter EC measured under Health Canada Intense smoking machine conditions.

Figure 8 shows Table 10, setting out the sidestream smoke yields under ISO smoking machine conditions Figuie 9 shows a comparison of HCI machine toxicant yields from ECs (lmg ISO) with those from published data sources. Figure 10 shows a comparison of HCI machine toxicant yields from ECs (6mg ISO) with those from published data sources.

Figure 11 shows a comparison of Total Toxicant Yields (TTY) between EC yields and published HCI yield data. Figure 12 shows a comparison of total yields from a subset of toxicants (TSY) between EC yields and published HCI yield data.

Figuie 13 shows a comparison of total normalised toxicant yields (NTT) between ECs and published HCI yield data.

Figure 14 shows a summary of the process by which high activity polymer-derived carbon is prepared.

Figure 15 shows Table 15, setting out the smoke and biomarker changes for test products as compared with a control cigarette.

Figure 16 shows the in vivo study design.

Figures 17 and 18 show the results of the in vivo study. Figure 19 shows a smoking article design according to an embodiment of the invention.

Two low toxicant tobacco blends, featuring a tobacco substitute sheet (TSS) or a tobacco blend treatment (BT), were combined with filters containing an amine functionalised resin material (CR20L) and/ or a high activity carbon adsorbent (HAC) to generate three experimental cigarettes (ECs). Mainstream smoke (MS) yields of smoke constituents were determined under four different smoking machine conditions. Health Canada Intense (HCI) machine smoking conditions gave the highest MS yields for nicotme-free dry particulate matter and for most smoke constituents measured. Constituent yields from the ECs were compared with those from two commercial comparator (CC) cigarettes, three scientific control (SC) cigarettes and published data on 120 commercial cigarettes. The ECs were found to generate some of the lowest machine yields of constituents from cigarettes for which HCI smoke chemistry is available; these comparisons therefore confirm that the ECs generate reduced MS machine constituent yields in comparison to commercial cigarettes.

The first stage in the design of a cigarette-based PREP involved the development of technologies which reduce the yields of smoke constituents. Experimental cigarettes (ECs) were assembled using these technologies and then assessed for their constituent yields using smoking machines; comparison to relevant control and reference products indicated the effectiveness of the cigarette design in generating reduced yields of constituents. Those ECs which are found to reduce smoking machine measured yields of smoke constituents, m comparison to reference products, are termed "reduced machine-yield prototypes" (RMYPs).

The inventors have described different individual technological approaches to the reduction of constituents m cigarette smoke, one of which involves the selection of tobacco blend components to provide a blend with reduced levels of the known precursors of undesirable smoke constituents, two of which modify the tobacco and two of which modify the cigarette filter. The tobacco blend (TB), the tobacco- substitute sheet material (TSS) and the tobacco blend treatment (BT) reduce the generation of constituents at source within the burning cigarette. The two filter technologies, an amine functionalised resin material (CR20L) and a high activity, polymer-derived, carbon adsorbent (HAC), remove volatile species from the smoke stream after formation. These technologies are discussed m greater detail below (m Section 2.1).

Tobacco Blend

This involves the selection of tobacco blend components that exhibit low levels of the precursors of undesirable smoke constituents, such as TSNAs and metals. For example, the levels of TSNAs may be reduced by using specific (such as lighter) tobacco blends and by selecting parts of the tobacco plant that are low m nitrate, a precursor of TSNAs. The person skilled in the art would be well aware of the ways in which the blending process may be adapted to provide a tobacco blend having these desired properties.

The tobacco blend may also comprise expanded tobacco, which is cut tobacco that has been expanded to reduce the mass of tobacco burnt m a cigarette. The expansion processes are similar to those used to make puffed rice snack food. One process used is called dry-ice expanded tobacco (DIET) and involves permeating the tobacco leaf structure with liquid carbon dioxide before warming. The resulting carbon dioxide gas forces the tobacco to expand. Some of the commercially available tobacco brands with low ISO tar yields use some proportion of expanded tobacco in the overall blend.

Tobacco Blend Treatment

Treated tobacco blends are described herein which have been treated by processes that allow the removal of protein and polyphenols from tobacco, with a beneficial effect on the smoke toxicant yields. The tobacco treatment was carried out on cut, flue-cured tobacco, and involved extraction of the tobacco with water followed by treatment with an aqueous protease enzyme solution. After treatment of the tobacco extract with adsorbents and concentration, the solubles were re-applied to the extracted tobacco. The treated tobacco retained the structure of the original tobacco and was made into cigarettes using conventional cigarette malting equipment, without the need for reconstitution into a sheet material.

Tobacco Substitute Sheet

Another approach to reducing smoke toxicant yields is to dilute the smoke with glycerol and it is proposed to include up to 60% of a glycerol-containing "tobacco substitute" sheet in cigarettes. Analysis of mainstream smoke from such

experimental cigarettes showed reductions in yields of most measured constituents, other than some volatile species.

Amme-functionalised resin material

It has been found that chemisorption is capable of removing high volatility aldehydes and HCN from mainstream cigarette smoke. A weakly basic macroporous polystyrene resm cross-linked with divmyl benzene, with surface amine

functionality, was identified and assessed as a cigarette filter additive. The material, manufactured by Mitsubishi Chemical Corporation is normally supplied m bead form in an aqueous environment and sold under the trade name Diaion®CR20 (hereafter referred to as CR20). This material offers the potential for the nucleophilic capture of aldehydes from mainstream smoke, and due to its weakly basic nature it may also be used for the removal of HCN from MS.

The amine-functionalised chelating resin material may be incorporated into the filter of a smoking article in a cavity, or dispersed (dalmation style) throughout the filter material (such as cellulose acetate) in the whole or a section of the filter.

High Activity Carbon

A high activity material comprising spherical particles of polymer-derived carbon was prepared by a propriety process (Von Bhicher and De Ruiter 2004; Von Blucher et el 2006; Bohringer and Fichtner 2008) and was available from Blucher GmbH (Germany). The polymer-derived material is approximately twice as effective, in general, at removing volatile cigarette smoke toxicants than the coconut shell- derived carbon commonly used in contemporary carbon filtered cigarette products. The polymer-derived carbon performed well at both ISO and HCI smoking regimes and with regular and smaller circumference cigarettes. Limitations were also obseived under higher flow-rate smoking conditions m the removal of acetaldehyde.

The high activity carbon may be incorporated into the filter of a smoking article m a cavity, or dispersed (dalmation style) throughout the filter material (such as cellulose acetate) m the whole or a section of the filter.

The present invention provides ECs made using combinations of the blend and filter technologies described. The goal of the study of these ECs was to assess whether these technologies could be combined into prototypes which reduce machine yields of toxicants in comparison to commercial products, and have the potential to reduce exposure of smokers to toxicants m human smoking.

Testing the ECs under a variety of smoking machine conditions and analysing the yields of smoke constituents on a per cigarette basis and as a ratio per milligram of nicotine yield, permits comparisons with relevant commercial comparator cigarettes, and also to a wide range of products reported m the literature. The results presented in this work demonstrate that the development of combustible RMYPs is feasible. 2. Materials and Methods

2.1 Design of Experimental, Control and Compaiator Cigarettes

The ECs were constructed from combinations of blend and filter technologies that were developed to reduce specific chemical classes of smoke toxicants or their precursors m tobacco (Table 1). For each EC individual tobacco grades with low TSNA and metal contents were selected and blended to provide a low toxicant starting point for the design of experimental cigarettes.

Table 1. Technologies used in the construction of experimental cigarettes (ECs).

Tobacco Blend Treatment

Briefly, the tobacco blend is subjected to an aqueous extraction step and the extract is subsequently passed through two stages of filtration to remove polyphenols and soluble peptides. The residual tobacco solids are treated with protease to remove insoluble proteins. After washing and enzyme deactivation, the tobacco sohds and filtered aqueous extract are re-combined. The treatment process lesults in reduced smoke yields of plienolics, aromatic amines, HCN, and a number of other nittogenous smoke constituents; however, there are also increases m the yields of formaldehyde and isoprene.

The tobacco material to be extracted may be strip, cut, shredded or ground tobacco. In a preferred embodiment, the tobacco is shredded tobacco. Other forms of tobacco may, however, be extracted using the methods described herein.

The tobacco material may be mixed with a solvent for extraction to form a slurry. The solvent may be added to the tobacco material in a ratio of between 10:1 and 50:1, preferably between 20:1 and 40:1 and most preferably between 25:1 and 30:1 by weight. In a particularly preferred embodiment, the solvent is added to the tobacco material in a ratio of 27:1 by weight.

The solvent may be an organic solution, but preferably is an aqueous solution or is water. At the very start of the extraction process, the solvent is usually water, but it can also contain alcohols such as ethanol or methanol, or it can contain a surfactant. Other solvents could be used, depending on the particular constituents to be extracted from the tobacco.

The extraction may be performed at 15-85°C, and preferably is performed at 65°C. It is preferable for the slurry to be continually stirred during extraction, such that the tobacco remains m suspension. Extraction should be performed for between 15 minutes and two hours. In a preferred embodiment, extraction is performed for approximately 20 minutes.

During extraction, soluble tobacco components are removed from the tobacco material and enter solution. These include nicotine, sugars, some proteins, ammo acids, pectins, polyphenols and flavours. Up to about 55% of the initial tobacco weight may become solubilised. It is important that the pectins m the tobacco fibre remain cross-linked throughout the extraction and treatment process m order to maintain the fibrous structure of the tobacco. Accordingly, calcium may be added to the solvent used to extract the tobacco and to any solutions used m the downstream processing procedures. Following extraction, the slurry may be drained to allow the liquid filtrate (the "mother filtrate") to be collected. Meanwhile, the insoluble tobacco residue may be further extracted by counter-current washing as it is conveyed, so that as many soluble constituents as possible are removed from the tobacco.

Fresh solvent may be applied to the tobacco and the filtiate (the "wash filtrate") is collected. The wash filtrate may be recycled by being applied to the incoming tobacco residue travelling on the belt at an upstream point. The collection and upstream reapplication of wash filtrate to incoming tobacco residue may be repeated a number of times, preferably three, four or even five times. Thus, the final wash filtrate that is collected at the head of the belt may be concentrated in those soluble tobacco constituents that have been removed from the tobacco residue as it travels the length of the filter. The final wash filtrate may be further recycled by being added to fresh tobacco to form a tobacco slurry, ready for extraction. For example, the final wash filtrate may be added into the tobacco mix tank where a tobacco slurry is formed prior to extraction. The extraction process may thus be a continual process in which fresh tobacco is extracted using recycled wash filtrate. Only at start-up of this extraction process is tobacco extracted with fresh solvent. Once the extraction process has begun, no fresh solvent is used m the extraction, but the solvent is solely made up of recycled wash filtrate.

As the extraction process continues, the extract thus becomes more concentrated in soluble tobacco constituents. These constituents include those that entered solution during primary extraction m the extraction tank (forming the mother filtrate), as well as those that entered solution during secondary extraction on the horizontal belt filter (forming the wash filtrate).

The final filtrate thus comprises both the mother and wash filtrates. In so doing, the tobacco residue that results after filtration is devoid of those constituents that are soluble in the solvent used for extraction. The extracted tobacco may be squeezed at the end of filtration, so as to remove any excess liquid from it. The extracted tobacco emanating from the horizontal belt filter is thus typically m the form of a dewatered mat. The final filtrate, hereinafter referred to as the tobacco extract, may be subsequently processed to remove those constituents not desired in the final tobacco product. Undesirable constituents include proteins, polypeptides, amino acids, polyphenols, nitrates, amines, nitrosamines and pigment compounds. The levels of constituents which may be considered desirable, such as sugar and nicotine, may, however, remain unaffected so that the flavour and smoking properties of the extracted tobacco are comparable to those of the original material.

In a preferred embodiment, the tobacco extract is treated to remove proteins, polypeptides and/or amino acids. Up to 60% of the proteins contained in the original tobacco material may be removed using an insoluble adsorbent such as hydroxyapatite or a Fuller's Earth mineral such as attapulgite or bentonite. The tobacco extract is preferably treated with bentonite, to remove polypeptides therefrom. Bentonite may be added to the extract in an amount of 2-4% of the weight of tobacco initially extracted. Alternatively, the tobacco extract may be fed into a tank containing a slurry of bentonite in water. A suitable slurry contains approximately 7 kg of bentonite in approximately 64 kg water (quantities per hour), for example, 7.13 kg bentonite m 64.18 kg water (quantities per hour). In any case, the bentonite concentration should be high enough to substantially reduce the protein content of the tobacco extract, but not so high as to additionally adsorb nicotine from it. Bentonite treatment may also be effective in the removal of pigment compounds found m tobacco extract which, if not removed, tend to darken the extract after concentration. When sufficient bentonite is used to treat the extract, the reduced amount of pigment compounds may result in a product that is not overly darkened m appearance. Following bentonite treatment, the tobacco extract may be purified from the slurry by centrifugation and/ or filtration. The tobacco extract may also, or alternatively, be treated to remove polyphenols therefrom.

Polyvinylpolypyrrolidone (PVPP) is an insoluble adsorbent for polyphenols, traditionally used m the brewing industry to remove polyphenols from beer. PVPP in an amount of 5-10% of the weight of tobacco initially extracted may be added to the extract. This amount of PVPP is capable of removing between 50 and 90% of the polyphenols in solution. The optimum pH for removal of polyphenols from the tobacco extract by PVPP is believed to be about 3. The efficiency of adsorption by PVPP may therefore be increased by reducing the pH of the extract via the addition of a suitable acid, such as hydrochloric acid. As an alternative to using PVPP to adsorb the polyphenols, one or more enzymes may be added to the tobacco extract to degrade the polyphenols therein. A suitable enzyme is laccase (urishiol oxidase). The invention is not, however, hmited to methods for removing only proteins and/or polyphenols from tobacco. Alternative or additional enzymes, agents or adsorbents may be used to remove other undesirable tobacco constituents from the tobacco extract. Examples of further undesirable tobacco constituents that could be removed from the extract include nitrates, amines and nitrosamines.

If a plurality of constituents is to be removed from the tobacco extract, a number of tanks may be set up m series, each one comprising a different enzyme, agent or adsorbent, m order for a chosen complement of undesirable constituents to be removed. Alternatively, a single tank may contain a plurality of enzymes, agents or adsorbents so that the undesirable constituents may be removed in a single step. For example, a bentonite or PVPP holding tank could comprise one or more additional enzymes, agents or adsorbents so as to remove not only protein or phenols from the tobacco, but one or more further undesirable constituents also.

Following treatment of the tobacco extract to remove the selected undesirable constituents, the extract is preferably concentrated to a solids concentration of between 20 and 50% by weight. Concentrations of up to 10% solids are most efficiently achieved using reverse osmosis. A further concentration to

approximately 40% solids may be achieved by means of a falhng film evaporator. Othei methods of concentration can be used and will be known to a person skilled m the art. The concentrated tobacco extract may be subsequently recombined with the extracted tobacco.

The tobacco, having been extracted in an aqueous solution as discussed above, however, is preferably further extracted to remove one or more further undesirable constituents before being recombined with the concentrated tobacco extract. Further extraction of the tobacco may be performed using an enzyme specifically selected for removal of the constituent of choice. In a preferred embodiment, the enzyme is a proteolytic enzyme for removal of protein from the tobacco. The enzyme is preferably a bacterial or fungal enzyme and, more preferably, is an enzyme used commercially in the food and detergent industries. The enzyme may be selected from the gioup consisting of Savinase™, Neutrase™, Enzobake™ and Alcalase™, which are all available from Novozymes A/S. The proteolytic enzyme is preferably added to the tobacco in an amount of between 0.1 and 5% by weight of the tobacco material. For example, Savinase™ may be added to the tobacco m an amount of approximately 1% by weight. The tobacco may be reslurried in a solution of the chosen enzyme. The ratio of water to tobacco in the slurry should be between 10:1 and 50:1, preferably between 20:1 and 40:1 and most preferably between 25:1 and 30:1 by weight. In a particularly preferred embodiment, the ratio of water to tobacco is 27:1 by weight. The pH of the tobacco/ enzyme mixture should be that which promotes optimal enzyme activity. Accordingly, it may prove convenient to feed the dewatered mat of tobacco into a tank m which the pH is adjusted, for example, by the addition of a base such as sodium hydroxide. The pH-adjusted tobacco may then be fed into an enzyme dosing tank for mixing with the enzyme of choice. The tobacco/ enzyme mixture may subsequently be fed into a plug flow reactor, where the enzymic extraction is performed. The enzymic extraction should be carried out at the temperature promoting optimal enzyme activity. Preferably, a narrow temperature range, such as 30-40°C, should be used to avoid denaturing the enzyme. The optimum working conditions when Savinase™ is the chosen enzyme are 57 °C and pH 9-11. The enzymic extraction should be carried out for at least 45 minutes; any shorter duration is believed to be insufficient for a proteolytic enzyme to degrade tobacco proteins.

Of course, multiple enzymic extractions could be carried out if there are multiple constituents to be removed from the tobacco. These could be performed in series or multiple enzymes could be added to the tobacco m a single treatment step. It also remains possible for the enzyme to be included in the very first extraction step in the treatment process, rather than forming a subsequent separate extraction step.

Following enzymic extraction, the insoluble tobacco residue may be washed with a salt solution, preferably a sodium chloride solution, to rinse it free of enzyme. Salt rinsing may be performed in a sequential, counter-current fashion.

Salt and water rinsing, however, may not be sufficient to remove all of the enzyme from the tobacco. The washed tobacco may also be treated to deactivate any residual enzyme remaining in the tobacco following the salt and water rinses. This may be done by steam treating the tobacco sufficiently to deactivate the enzyme, but not so much that the tobacco loses its fibrous form. In an embodiment, steam treating is carried out at 98°C for four minutes, but the residence time may be increased to 10 minutes or so if desired. Alternatively, the tobacco may be heat treated to deactivate the enzyme, for example by microwavmg or baking the tobacco. In another embodiment, the enzyme may be deactivated by chemical denaturation; steps should however be taken to remove the chemical from the tobacco.

The processed tobacco may then be recombmed with the concentrated tobacco extract. Adding the treated extract back to the extracted tobacco ensures retention of water soluble flavour components of tobacco and nicotine in the final product. Recombination therefore results m a tobacco product that has similar physical form and appearance, taste and smoking pioperties to the original material, but with substantially reduced levels of protein, polyphenols or other constituent(s) of choice. Recombination may be achieved by spraying the tobacco extract onto the tobacco. The amount of the original extract being recombined with the processed tobacco depends upon the amount that was lost during treatment of the extract to remove selected constituents, and will vary from one type of tobacco to the next.

A standard diying process may be used to dry the treated tobacco, either before, during or after recombination with the treated tobacco extiact. The starting moisture content of the treated tobacco is typically approximately 70-80%. In a preferred embodiment, the moisture content after drying should be approximately 14%. A heated dryer, such as an apron dryer, may be used to reduce the starting moisture content in the tobacco to approximately 30%. A second heated dryer, such as an air dryer, may then be used to further reduce the moisture content to approximately 14%.

The final dried product may subsequently be processed into a finished form, such as a sheet, which, when shredded, can form all or part of a cigarette filler. Owing to as much as 30% of the original constituents of tobacco being removed therefrom during the extraction and treatment process, however, the concentration of remaining constituents per unit weight of tobacco is increased in the finished product compared to the original material. These constituents include cellulose, which, together with sugars and starches, may produce harmful volatile materials such as acetaldehyde and formaldehyde in smoke when combusted.

Tobacco Substitute Sheet

Incorporation of the tobacco substitute sheet (TSS) into a tobacco blend reduces the quantity of tobacco in a cigarette, thereby diminishing the overall potential for the cigarette to generate toxicants. The TSS also contains glycerol and, when heated, the TSS releases glycerol into the smoke stream contributing to the total amount of particulate smoke, measured as nicotme-free dry particulate matter (NFDPM, also known as "tar"). As most cigarettes are designed to meet a specific NFDPM yield value, incoiporation of glycerol into the smoke stream effectively results in a reduced contribution of the tobacco combustion products to the overall NFDPM value: this process is termed "dilution." The incorporation of TSS into cigarettes results in reductions in a wide range of smoke constituents, including both particulate and vapour phase toxicants. In vitro toxicological tests showed reductions in the activity of smoke particulates in proportion to their glycerol content. Human exposure to nicotine was reduced by a mean of 18% as determined by filter studies and by 14% using 24 hour urinary biomarker analysis. Smoke particulate exposures were reduced by a mean of 29% in filter studies and by similar amounts based on urinary 4-(methylnitrosamino)-l-(3- pyridyl)-l-butanol

concentrations. These results show that reducing exposure to some smoke toxicants is possible using a tobacco substitute sheet. According to the present invention, a smoking article may be prepared including a tobacco substitute sheet material comprising a non-combustible inorganic filler material, an alginic binder and aerosol generating means.

Advantageously the tobacco substitute sheet material comprises as the main components thereof, non-combustible inorganic filler, binder and aerosol generating means, with these three components together preferably comprising at least 85% by weight of the tobacco substitute sheet material, preferably greater than 90%, and even more preferably total about 94% or more by weight of the tobacco substitute sheet material. The three components may even be 100% of the tobacco substitute sheet material. The remaining components are preferably one or more of colourant, fibre, such as wood pulp, or flavourant, for example. Other minor component materials will be known to the skilled man. The tobacco substitute sheet material is therefore a very simple sheet in terms of its constituents.

As used herein, the term ' tobacco substitute sheet material' means a material which can be used in a smoking article. It does not necessarily mean that the material itself will necessarily sustain combustion. The tobacco substitute sheet material is usually produced as a sheet, then cut. The tobacco substitute sheet material may then be blended with other materials to produce a smokeable filler material.

The present invention further provides a smoking article comprising a wrapped rod of a smokeable filler material, the smokeable filler material consisting of a blend which incorporates tobacco substitute sheet mateiial comprising a non-combustible inorganic filler, an alginic binder and aerosol generating means, the smoking article having an aerosol transfer efficiency ratio of greater than 4.0. As used herein, the aerosol tiansfer efficiency is measured as the percentage aerosol m the smoke divided by the percentage aerosol in the smokeable filler material. Preferably the aerosol tiansfer efficiency is greater than 5, and more preferably greater than 6.

The smokeable filler material used in the smoking articles of the present invention may comprise a blend consisting of not more than 75% by weight of the tobacco substitute sheet material. Preferably the inorganic filler material is present in the range of 60-90%, and is more preferably greater than 70% of the final sheet material. Advantageously the inorganic filler material is present at about 74% by weight of the final sheet material, but may be present at higher levels, for example, 80%, 85% or 90% by weight of the final sheet material.

The non-combustible filler advantageously comprises a proportion of material having a mean particle size m the range of 500μιη to 75μηι. Preferably the mean particle size of the inorganic filler is in the range of 400μηι to ΙΟΟμιη, and is more than 125μιη, and preferably more than 150μιη. Advantageously the mean particle size is at or about 170μηι, and may be in the range of 170μιη to 200μηι. This particle size is in contrast to that conventionally used for food grade inorganic filler materials m alternative tobacco products, namely a particle size of about 2-3 μηι. The lange of particle size seen for each inorganic filler individually may be from Ιμΐη-lmm (ΙΟΟΟμηι). The inorganic filler material may be ground, milled or precipitated to the desired particle size.

Advantageously the inorganic filler material is one or more of perlite, alumina, diatomaceous earth, calcium carbonate (chalk), vermicuhte, magnesium oxide, magnesium sulphate, zinc oxide, calcium sulphate (gypsum), ferric oxide, pumice, titanium dioxide, calcium aluminate or other insoluble aluminates, or other inorganic filler materials. The density range of the materials is suitably m the range of 0.1 to 5.7 g/cm³. Advantageously, the inorganic filler material has a density that is less than 3 g/cm³, and preferably less than 2.5 g/cm³, more preferably less than 2.0 g/cm³ and even more preferably less than 1.5 g/cm³. An inorganic filler having a density of less than 1 g/cm³ is desirable. A lower density inorganic filler reduces the density of the product, thus improving the ash characteristics.

If a combination of inorganic filler materials is used, one or more of the fillers may suitably be of a small particle size and another may be of a larger particle size, the proportions of each filler being suitable to achieve the desired mean particle size. The static burn rate required m the finished smoking article may be achieved using an appropriate blend of tobacco and tobacco substitute sheet material m the smokeable filler material. Pf eferably the inorganic filler material is not m agglomerated form. The inorganic filler material should require little pre-treatment, other than perhaps size gradation, before use. Preferably the binder is present in the range of about 5-13%, more preferably less than 10% and even more preferably less than 8%, by weight of the final filler material. Advantageously the binder is about 7.5% by weight or less of the final sheet material. Advantageously, if the binder is a mixture of alginate and non-algmate binders, then preferably the binder is comprised of at least 50% alginate, preferably at least 60% alginate and even more preferably at least 70% alginate. The amount of combined binder requited may suitably decrease when a non-alginate binder is utilised. The amount of alginate in a binder combination advantageously increases as the amount of combined binder decreases. Suitable algmic binders include soluble alginates, such as ammonium alginate, sodium alginate, sodium calcium alginate, calcium ammonium alginate, potassium alginate, magnesium alginate, triethanol-amine alginate and propylene glycol alginate. Other organic binders such as cellulosic binders, gums or gels can also be used m combination with algmic binders. Suitable cellulosic binders include cellulose and cellulose derivatives, such as sodium carboxymethylcellulose, methyl cellulose, hydroxypropyl cellulose, hydroxyethyl cellulose or cellulose ethers. Suitable gums include gum arable, gum ghatti, gum tragacanth, Karaya, locust bean, acacia, guar, quince seed or xanthan gums. Suitable gels include agar, agarose, canageenans, furoidan and furcellaran. Starches can also be used as organic binders. Other suitable gums can be selected by reference to handbooks, such as Industrial Gums, E. Whistler (Academic Press). Much preferred as the major proportion of the binder- are alginic binders. Alginates are preferred in the invention for their neutral taste character upon combustion.

Preferably the aerosol generating means is present m the range of 5-20%, more preferably is less than 15%, is even more preferably greater than 7% and even more preferably is greater than 10%. Preferably the aerosol generating means is less than 13%. Most preferably the aerosol generating means is between 11% and 13%, and may advantageously be about 11.25% or 12.5%, by weight of the final sheet material. Suitably the amount of aerosol generating means is selected in

combination with the amount of tobacco material to be present m the blend comprising the smokeable fillet material of a smoking article. For example, in a blend comprising a high proportion of sheet material with a low proportion of tobacco material, the sheet material may require a lower loading level of aerosol generating means therein. Alternatively in a blend comprising a low proportion of sheet material with a high proportion of tobacco material, the sheet material may require a higher loading level of aerosol generating means therein.

Suitable aerosol generating means include aerosol forming means selected from polyhydric alcohols, such as glycerol, propylene glycol and tnethylene glycol; esters, such as triethyl citrate or triacetin, high boihng point hydrocarbons, or non-polyols, such as glycols, sorbitol or lactic acid, foi example. A combination of aerosol generating means may be used.

An additional function of the aerosol generating means is the plasticismg of the sheet material. Suitable additional plasticisers include water. The sheet material may suitably be aerated. The cast slurry thereby forms a sheet material with a cellular structure.

Advantageously the or a proportion of the aerosol generating means may be encapsulated, preferably micro-encapsulated, or stabihsed in some other way. In such cases the amount of aerosol generating means may be higher than the range given. Advantageously the smoking material comprises a colourant to darken the material and/ or a flavourant to impart a particular flavour. Suitable flavouring or colourant materials, subject to local regulations, can include cocoa, liquorice, caramel, chocolate or toffee, for example. Finely ground, granulated or homogenised tobacco may also be used. Industry approved food colorants may also be used, such as El 50a (caramel), El 51 (brilliant black BN), El 53 (vegetable carbon) or El 55

(brown HT). Suitable flavourants include menthol and vanillin, for example. Other casing materials may also be suitable. In the alternative, the presence of vermicuhte or other inorganic filler materials may give a darker colour to the tobacco substitute sheet material. Preferably the colourant is present from 0-10% and may be as much as 5-7% by weight of the final tobacco substitute sheet material. Advantageously the colourant is less than 7%, preferably less than 6% and more preferably less than 5% of the final tobacco substitute sheet material. Much preferred is use of colourant at less than 4%, less than 3% and less than 2%. Cocoa may suitably be present in a range of 0-5% and liquorice may be present m a range of 0-4%, by weight of the final tobacco substitute sheet material. When the colourant is cocoa or liquorice, for example, the minimum amount of cocoa to obtain the desired sheet colour is about 3% and for liquorice is about 2%, by weight of the final tobacco substitute sheet material. Similarly, caramel may suitably be present m a range of 0-5%, preferably less than about 2% by weight of the final tobacco substitute sheet material, and more preferably about 1.5%. Other suitable colorants include molasses, malt extract, coffee extract, tea resinoids, St. John's Bread, prune extract oi tobacco extract. Mixtures of colorants may also be used.

If permitted under local regulations, flavourants may also be added to alter the taste and flavour characteristics of the tobacco substitute sheet material. Advantageously, if a food dye is utilised in the alternative it is present at 0.5% by weight or less of the final tobacco substitute sheet material. The colourant may alternatively be dusted into the sheet after sheet manufacture.

Fibres, such as cellulose fibres, for example wood pulp, flax, hemp or bast could be added to provide the sheet material with one or more of a higher strength, lower density or higher fill value. Fibres, if added, may be present m the range of 0.5-10%, preferably less than 5% and even more preferably less than about 3% by weight of the final sheet material. Advantageously there is no fibrous material present in the sheet material, cellulosic or otherwise.

Advantageously the tobacco substitute sheet material is a non-tobacco containing sheet. It shall be understood that at high levels of sheet material inclusion in the blend, e.g. at greater than 75% by weight of the blend, the combustibility of the blend is poor. This may be overcome by, for example, incorporating low levels of up to 5-10% granular carbon m the tobacco substitute sheet material. The carbon is prefeiably not an agglomerated carbonaceous material, i.e. the carbon is not pre- treated by mixing with another material to produce an agglomerate. Preferably the tobacco substitute sheet material is blended with tobacco material to provide smokeable filler material. Preferably the tobacco material components in the blend are high quality lamina grades. Advantageously the majority of the tobacco material is cut tobacco. The tobacco material may comprise between 20- 100% expanded tobacco of a high order expansion process, such as DIET for example. The filling power of such material is typically in the range of 6- 9cc/g (see GB 1484536 or US 4,340, 073 for example).

Preferably the blend comprises < 30% of other blend components apart from lamina, the other blend components being stem cut rolled stem (CRS), water treated stem (WTS) or steam treated stem (STS) or reconstituted tobacco. Preferably the other components comprise < 20%, more preferably < 10% and even more preferably < 5% of the final weight of the tobacco material.

Suitably a smoking article according to the invention comprises tobacco material being treated with aerosol generating means. The tobacco material may be treated with aerosol generating means, but this is not essential for all blends of tobacco material and sheet material. The amount of aerosol generating means added to the tobacco is m the range of 2- 6% by weight of the tobacco. The total amount of aerosol generating means m the blend of tobacco material and sheet material after processing is advantageously in the range of 4-12% by weight of the smokeable material, preferably less than 10% and preferably more than 5%. High Activity Carbon

The polymer-derived, high activity carbon granules used in the dual and triple stage filters possesses a pore structure different from the carbon commonly used in commercial cigarettes, which is typically derived from coconut shells. As a result it has superior adsorption characteristics for a range of volatile smoke toxicants. The spherical particle shape polymer-derived carbon was prepared by a propriety process (Von Blucher and De Ruiter 2004, Von Blucher et el 2006, Bohrmger and Fichtner 2008), as depicted in Figure 14. The polymer-derived active carbon is produced using a batch process with indirect heated rotary kilns, under reduced pressure in an melt atmosphere. After preparation of the spherical polymer feed- stock the material is thermally stabilised using an excess of oleum. Subsequently, the mateml is slowly heated to 500°C, resulting in the release of predominantly S0₂ and H₂0 and the carbonisation of the polymer. The resulting carbon has an initial pore system which is not accessible for typical adsorptives. To create a porous system capable for adsorption, the material is further heated to 900 to 1000°C for activation with oxidising agents (steam). This establishes a pole system consisting mainly of micropores with pore sizes between 0.7 and 3 nm. Subsequent activation with C0₂ leads to the formation of predominantly larger mesopores in the range of 3 to 80 nm. Combining the steam and C0₂ activation steps offers a flexible strategy for producing desired pore characteristics. The polymer-derived carbon, being a synthetic material, possesses a much more closely defined spherical shape, together with a more uniform particle size. The polymer derived material possesses a lower density, and has a lower ash content reflecting the synthetic nature of the polymer feedstock in comparison to a natural coconut shell as starting materials for the carbonization processes. Most smoke constituents are adsorbed more effectively by the polymer-derived carbon under the ISO regime than by activated coconut carbon, with reductions of the order of 80-95% observed with smoke constituents other than formaldehyde, acetaldehyde, hydrogen cyanide (HCN) and toluene (50-60% reductions). Under HCI conditions, cigarettes with conventional coconut carbon provide reductions of the order of 25-45% for most smoke constituents, other than acetaldehyde (16%). The cigarettes including polymer-derived carbon reduce most smoke constituent yields by 60-90%, other than acetaldehyde and HCN (15-30%).

Amme-functionalised Resin Beads

DIAION® CR20 is a commercially available type of amme-functionalised resin bead which may be used m the present invention (manufactured by Mitsubishi Chemical Corporation). It has polyamine groups as chelating ligands which are bonded onto a highly porous crosslmked polystyrene matrix. CR20 shows large affinity for transition metal ions. The exact type of amine gioups produced by functionahzation cannot be precisely controlled and several different types could be present on the resins. Commercial grade CR20 (hereafter referred to as CR20C) was found to have a characteristic odour incompatible with conventional consumer acceptable cigarette smoke character when incorporated into cigarettes. However, modification to the synthesis conditions by Mitsubishi significantly reduced the intensity of this odour, resulting in a "low-odour" grade of CR20 (hereafter referred to as CR20L). In this work, unless otherwise stated, all results obtained refer to CR20L. This material possessed a bead size of 600mm, density of 0.64 g/ cm³, a 15% by weight water content, and total exchange capacity of 0.92 meq/cm³.

Various other types of CR20 are made by Mitsubishi Chemical Corporation, including CR20D and CR20HD. All of the different types or grades of the ion- exchange resin are encompassed by the term CR20 as used herein.

Some CR20 beads are provided m water and, to make them suitable for use in a cigarette filter application, it may be necessary to remove at least some of the water. In one embodiment, the water is removed and the material is dried to approximately 15% or less moisture. In an alternative embodiment, a higher moisture content may be acceptable m the filter of smoking articles.

CR20, including specifically CR20L, may be incorporated into cigarette filters. In comparison to filters containing conventional carbon, CR20L offers superior reductions for HCN, formaldehyde and acetaldehyde. However, carbon is more efficient than CR20L in removing other volatile constituents from a smoke stream.

Experimental Cigarettes

Cigarettes were constructed using these technologies targeting ISO NFDPM (tar) yields of 1 and 6mg.

Three scientific control cigarettes were also manufactured to allow an evaluation to be made of the contribution of the filter technologies to smoke constituent reductions from ECs. Two commercial comparator cigarettes, a lmg ISO design and a 6mg ISO design, were also used m these studies. Comparisons with

commercial brands were conducted because realistic control cigarettes are required to assess the success with which the different smoke constituent reduction technologies can be brought together into a coherent and consumer acceptable cigarette design. Also, the use of commercial cigarettes allows examination of the extent with which constituent reductions can be realised against real-world cigarettes, rather than scientific controls. Finally, use of commercial reference products allows relevant comparisons to be made of sensory acceptability and human exposure under real-world use.

The commercial comparator products were of similar machine smoked constituent yields to the market leading brands at 1 mg and 6 mg (ISO) from Germany in 2007- 8. BAT group comparator cigarettes were chosen, rather than the actual market leading brands, m order that full information was available on blend and cigarette design characteristics, and to allow product masking to be conducted for human sensory and exposure evaluations. Samples of both commercial cigarettes were therefore manufactured specially for these studies, without brand marking or other identification, in order to support human smoking studies.

2.2 Specifications for Experimental, Comparator and Control Cigarettes

Common features were used in the design of the ECs: all were constructed to the same basic dimensions, of 84 mm cigarette length (a 57 mm tobacco rod plus a 27 mm filter), 24.6 mm circumference and the filters were all based on cellulose acetate (CA) fibres plasticized with triethyl citrate. Tobacco grades with low TSNA and metal contents were identified and combined for the tobacco blends used in these prototypes. Three different experimental cigarettes were prepared, and the design features of the three ECs are summarised and compared with control cigarettes and commercial comparators m Table 2 (shown in Figure 1) and described below.

The experimental cigarette BT1, combined a Virginia style tobacco blend containing BT treated tobacco (75.4% treated Virginia tobacco, with 4.3% Oriental tobacco and 20.3% untreated Virginia tobacco) with a filter containing a CR20 stage (to reduce formaldehyde, acetaldehyde and HCN yields) and a polymer-derived, high activity carbon filter containing stage (to reduce yields of isoprene and other volatile toxicants). The target NFDPM yield from this cigarette was 1 mg under ISO machine smoking conditions. The experimental cigarette TSS1 was also designed to yield 1 mg of NFDPM under ISO smoking machine conditions and was based on an US style blend containing TSS (a blend of Viiginia, Burley and Oriental tobaccos, with the inclusion of approximately 20% TSS and the same filter used in

experimental cigarette BTl. The experimental cigarette TSS6 also used 20%TSS in a different US style blend, and was designed to give an NFDPM yield of 6 mg under ISO machine smoking conditions. A different filter construction was used with this cigarette: a dual segment filter containing 80 mg of the high activity carbon interspersed amongst CA fibres adjacent to the tobacco rod with a CA stage at the mouth end.

The commercial comparator cigarette CCl contained a US-blended style of tobacco, including some Maryland tobacco. The commercial comparator cigarette, CC6, was also a typical US-blended cigarette but with a different blend to CCl. The design features of the three ECs are summarised and compared with control cigarettes and commercial comparators in Table 2 [shown m Figure 1]. Both commercial comparator cigarettes used single stage cellulose acetate filters. The three "scientific control" (SC) cigarettes had identical construction to the relevant experimental cigarettes BTl, TSSl and TSS6, with the exception that the filter used in each control cigarette was a single stage 27 mm CA filter without additional filter adsorbent media.

Table 2 shows that the cigaiette constructions of BTl and CCl were very similar, with well matched filter ventilation and paper permeability. There were slight differences in tobacco density and filter pressure drop (the draw resistance or impedence to flow of the filter), with BTl higher than CCl for both parameters. The cigarette constructions of TSSl and CCl were also very similar. The filter pressure drop was higher from TSSl than the commercial control, but both tobacco density and filter pressure drop were higher for CCl. For TSS6 and CC6 less filter ventilation was used than with the lmg (ISO) products. Comparing the two 6 mg (ISO) products showed slightly higher tobacco densities, pressure drop values and shghtly lower filter ventilation for TSS6.

2.3 Smoke Chemistry Analysis

Prior to smoke chemistry analysis, cigarettes were conditioned according to the specifications of ISO 3402, 1999. Routine chemical analyses were performed according to the smoking conditions specified in ISO 4387, 2000 (i.e., a 35 ml puff of 2 seconds duration taken every 60 seconds, abbreviated as 35/2/ 60) and ISO 3308, 2000 which was developed for NFDPM and nicotine analysis.

Approximately 150 smoke constituents have been described as toxicants and a few regulatory authorities have requested yield data on a subset (approximately 40) of them. Yield restrictions for some of these toxicants have been proposed (Burns, D., et al (2008) Mandated lowering of toxicants m cigarette smoke: a description of the World Health Organization TobReg proposal. Tob. Control 17, 132-141) along with an approach to their biomonitormg (Hecht, S.S. et al (2010) Applying tobacco carcinogen and toxicant biomarkers m product regulation and cancer prevention. Chem. Res. Toxicol. 23, 1001-1008). For these reasons and in order to characterise the ECs more precisely, the MS yields of an extended range (47 analytes) of smoke constituents were measured. The other, approximately 100, toxicants not examined in this work were not measured due to the lack of available validated analytical methods. Values for benzo(a)pyrene yields were obtained twice, through a diiect measure and also as part of a suite of polycyclic aromatic hydrocarbons (PAHs).

Slight modification to the ISO smoking parameters was required for the

measurement of other analytes, and the current methods are available from British American Tobacco, (www.batscience.com/groupms/ sites/BAT_7AWFH3.nsf/ vwPagesWebLive/D07AXLPY?opendocument&SKN=l). Measuring the yield of smoke constituents from a smoking machine does not mimic human smoking yields and so all RTPs were tested under a lange of different smoking machine settings in order to allow machine yield performance to be assessed over a wide range of possible smoking conditions. These modified smoking conditions are described in Table 3. Table 3. Smoking machine parameters.

Intense-Filter

Vents Open

ISO WG 9 WG9B 60 2 30 50

Intense Option B

Sidestream smoke (SS) yields were also measured as described by Health Canada, 1999 but only under ISO smoke generation parameters and for a wider range of smoke constituents. The SS testing was conducted by Labstat International ULC. 2.4 Statistical Analysis

Statistical comparisons of smoke yields between different cigarette types were conducted using a two-tailed, unpaired, Student's t-test, performed with Mimtab vl6. Levels of significance of P<0.01 and P<0.05 are shown and any P value >0.05 is shown as nonsignificant ( S). For comparisons of individual smoke constituent yields across studies, mean values from published data sets (Health Canada, 2004; Counts et al., 2005; Department of Health Australia, 2002) were examined for normal distribution using the Anderson Darling statistic. Percentile distributions within the toxicant data were calculated using an empirical cumulative distribution analysis within Minitab vl 6. 3. Results and Discussion

Testing of the ECs was conducted m order to examine the actual performance of the ECs from a blend and smoke chemistry perspective, by quantifying the MS constituents and specific toxicant yields under a number of machine smoking conditions. The SS emissions from the ECs were also measured using the ISO smoking profile. The tests were conducted on a comparative basis with two commercial cigarettes and with three scientific control cigarettes. As a final step, the overall performance of the ECs was assessed both m comparison to previously published MS yield data on cigarettes from several countries and as ratios of specific toxicant yields to nicotine yields.

3.1 Mainstream Smoke Constituent Yields

The yields of the major smoke constituents (NFDPM, nicotine and CO) and glycerol under four smoking machine conditions are shown in Table 4 (shown in Figure 2). Glycerol measurements are included m this table because it has been incorporated into the tobacco-substitute sheet used in the ECs TSSl and TSS6, to dilute other smoke constituents m the smoke particulate phase. Table 4 shows that BTl and CC1 were well matched across the four smoking regimes for MS NFDPM and nicotine yields, but that BTl had lower CO yields than CC1. TSSl and CC1 were well matched across the four smoking regimes for NFDPM and nicotine yields but TSSl had lower CO yields than CC1. The higher glycerol yield from TSSl is consistent with the intended dilution effect due to the glycerol content of TSS. The MS NFDPM and nicotine yields from TSS6 and CC6 were well matched across the four smoking regimes, other than higher CO yields from CC6 and the expected higher glycerol yields from TSS6.

For these major smoke analytes the yields measured followed the same rank order based on smoking machine conditions: ISO <HCI-VO <WG9B <HCI. The yield differences between the different regimes were substantially greater with the lmg products than with the 6mg products, as the level of ventilation was higher and the impact of ventilation blocking for the WG9B and HCI regimes is therefore more profound for the lmg products. For the 6mg products the differences m the major smoke measures (NFDPM, nicotine and CO) between some of the regimes were small (in the order of 5-10%).

The 47 toxicants quantified m this work were also measured under all of the smoking machine conditions shown in Table 3, except that data for the ECs TSSl and BTl under ISO machine smoking conditions were not collected because preliminary runs showed the yields of many constituents to be below the LOQ for the methods. The machine smoked yields of these toxicants generally followed the lank order noted for NFDPM, nicotine and CO shown in Table 4 and so, for the remainder of this paper, only the yields obtained under HCI conditions are described. Some consistent exceptions to the general yield trend were observed. With all products the volatile phenols, quinoline, and fluorene did not increase systematically with increasing intensity of the smoking regime and the yields of the major smoke measures; arsenic, phenanthrene and the measure for benzo(a)pyrene from the PAH suite also displayed this behaviour for the majority of the products. In particular the yields of these species were greater under the WG9B regime than with the HCI regime despite the greater overall amounts of smoke generated by the HCI regime. Volatile phenols are known to be selectively removed from smoke by cellulose acetate filters; the consistent behaviour observed here may represent some change in filtration efficiency for these species between the WG9B and HCI regimes. Alternatively it may represent some analytical weakness with the measurement method at high intensity smoking regimes. Similar changes were observed on a more occasional basis for some analytes (e.g. 1,3-butadiene yields with CC1 were lower than expected from the trends across smoking regimes found for the other five products); this was found in particular with the 6mg products when similar amounts of NFDPM were generated between the different smoking regimes, and these observations are likely due to analytical errors, or reflect limits m the discriminatory power of the analytical techniques.

The use of the HCI smoking regime in this work represents the strictest test of the ECs and the commercial comparator cigarettes. Although these smoking conditions inactivate a design feature used m the ECs and commercial cigarettes (filter ventilation), they address criticism of the machine yield values obtained from ventilated cigarettes.

3.1.1 Metal and TSNA Yields

Two groups of toxicants included on regulatory lists are the metals and the tobacco specific nitrosamines (TSNAs). Both these groups of toxicants are primarily affected by the tobacco blend used m cigarette manufacture and so careful blend selection is a major contributor to their reduction m smoke. The chemical analysis of blend metals and TSNAs are described m Table 5 (shown in Figure 3) and their MS yields under HCI smoking machine conditions are shown in Table 6 (shown m Figure 4). The yields are discussed for each EC m Sections 3.1.2.1 to 3.1.2.3 below.

3.1.2 Other Toxicant Yields

Smoke constituent yield comparisons between ECs and commercial controls, under HCI smoking machine conditions, are shown in Table 7 (shown m Figure 5). The yields are discussed for each EC in Sections 3.1.2.1 to 3.1.2.3 below. 3.1.2.1 BTl

Measurement of blend chemistries (Table 5) showed the blend arsenic and chromium contents of BTl were statistically significantly higher than the

commercial cigarette CCl; whereas lead and nickel contents of the BTl blend were lower. The MS yields for metals from BTl were comparable to or lower than the yields from CCl, except that the arsenic and mercury yield were higher. The higher arsenic yield may be explained by the higher blend content of this metal but the mercury yield is not explained by blend content and may represent an artefact because the BTl blend content of mercury was comparable to or lower than CCl, being below the LOQ for this metal (Table 5).

Blend nitrosamine content of BTl was lower than US-blended commercial comparator CCl, as has been seen previously m comparison of Virginia and US- blended cigarettes. The MS yields of nitrogenous constituents were expected to be lower from BTl than from CCl for two reasons: first the tobacco treatment reduces precursors of nitrogenous smoke compounds; and, second, Virginia style tobaccos typically generate lower yields of nitrogenous smoke constituents than US-blended cigarettes. Measurement of the yields of nitrogenous compounds showed the anticipated differences: yields of the TSNAs were statistically significantly (83-96%) lower from BTl than from CCl (Table 6); aromatic amine yields from BTl were 26- 57% lower than from CCl (Table 7); and the yields of other nitrogenous

compounds from BTl were also substantially lower (HCN by 82%, NO by 79%, ammonia by 75%, pyridine by 97%, quinolene by 67% and acrylonitnle by 69%) than the respective yields from CCl (Table 7). These data confirm that the blend selection, use of the BT process (and incorporation of CR20 m the filter in the case of HCN yields) produced the expected lower yields of toxicants from the EC.

The BT process also reduces blend polyphenol levels and so reductions in MS phenols yields would be expected; however, higher yields of phenohcs are generally expected from Virginia style products than from US-blended products and this tobacco type difference could mitigate any reductions from the BT process.

Comparison between phenolic compound yields from CCl and from BTl showed a mixed picture: phenol, p-cresol and lesorcinol yields were lower from BTl, whereas m-cresol, catechol and hydroquinone yields were higher from BTl (Table 7). The BT process does not influence benzo(a)pyrene yields and analysis of PAHs m the current study showed comparable yields from BTl and CCl for fluorene, phenanthrene, pyrene and benzo(a)pyrene. Lower carbonyl yields (26 to 74% lower) were obtained from cigarette BTl, apart from formaldehyde, which showed a higher (41%) yield from BTl . The volatile hydrocarbon yields from BTl were lower, with a range from 21 to 78% for isoprene, benzene, toluene and naphthalene, when compared to the respective constituent yields from CCl; however, the 1,3-butadiene yield was 35% higher from BTl compared to CCl . The 1,3-butadiene yields from CCl are lower than expected under the HCI regime, and this observation may therefore be unreliable. Most of the observed differences in volatile constituent yields are consistent with the use of an effective vapour phase adsorbent m the filter of BTl. Formaldehyde yields are driven m part by sugar levels, which are normally higher in Virginia blends than in US blends. Formaldehyde yields are also increased by the blend treatment process. Hence the higher formaldehyde yields from BTl are understandable on the basis of knowledge of formaldehyde generation m cigarettes. The apparent higher yield of 1,3-butadiene from BTl is possibly due to an error in the yield measurement of CCl as there is no obvious mechanistic factor to support this difference (the tobacco treatment process does not give statistically significant changes m 1,3-butadiene yields and the use of the vapour phase adsorbent m BTl filters should result in lower 1,3 butadiene yields from BTl). The contribution of the blend and the selective filter used in BTl to the overall reductions in smoke toxicants are addressed in Section 3.2 and the results are consistent with the higher yield values for formaldehyde observed m Table 7 being due to blend chemistry factors. 3.1.2.2 TSSl

The overall blend metal content was higher in TSSl than in CCl for some metals (arsenic, chromium and nickel), lower for cadmium content and not different for other metals (Table 5). The TSS contains a high proportion of chalk, which would contribute some portion of the blend metals. Analysis of the TSS showed a higher level of chromium and comparable or lower levels of the other measured metals than the TSSl blend. Hence, the higher chromium content of TSSl than CCl most likely reflects the inclusion of TSS material in the blend; whereas, the higher arsenic and nickel levels were most likely due to the different tobacco types used m the blend. It should be noted that the transfer of metals from the TSS would not necessarily occur with the same efficiency as from tobacco, due to possible differences in the chemical form (and therefore volatility) of trace metals in chalk and in tobacco. Thus, the metal yields in MS under HCI smoking machine conditions were either lower or not statistically significantly different when TSS1 was compared to CC1 (Table 6). The blend nitrosamme content of TSS1 was lower (23-72%) than that of CC1 (Table 5) and the MS yields of the TSNAs under HCI machine smoking conditions were correspondingly lower (17 to 69%) for TSS1 than CC1 (Table 6).

Statistically significantly lower yields were found from TSS1 than from CC1 for most of the phenohcs (29-57%), carbonyls (44-86%), PAHs (8 to 71%) and miscellaneous volatile constituents (27 to 94%); although for catechol,

hydioquinone and benzo(a)pyrene, these differences did not achieve statistical significance (Table 7). These data demonstrate lower toxicant yields from TSS1 across all of the analyte classes examined, and therefore support the expectation that the TSS, and the three stage filter, should function to give overall MS toxicant yield reductions m an EC.

3.1.2.3 TSS6

The blend metal contents of TSS6 and CC6 were similar, other than statistically significantly higher chromium and cadmium blend levels in TSS6. As noted above, the higher chromium level was most hkely due to the high inorganic content of the TSS; whereas, the higher cadmium content most likely reflects a difference in the tobacco types used between the two blends. The MS yields of cadmium and chromium, determined under HCI smoking machine conditions, were not elevated in TSS6 compared to CC6 (Table 6), which again supports the contention that the chemical form of these metals was different between the EC and the commercial comparator, and less likely to transfer into MS.

The blend nitrosamme contents were lower (39 to 54%) from TSS6 than those measured for the CC6 blend (Table 5). Again, this lower blend nitrosamme content translated to 37 to 50% lower MS yields for these TSNAs under HCI smoking machine conditions (Table 6).

The MS yields from TSS6, across all of the other chemical classes measured

(aromatic amines (13-20%), phenohcs (8-32%), carbonyls (35-85%), PAHs (18-81%) and miscellaneous volatile toxicants (41-96%)) were statistically significantly lower than the yields from CC6, except for 1- and 2-aminonaphthalene and m- and p- cresol where the values were not significantly different and for ammonia where the higher yield (13%) was not statistically significantly different to that of CC6 (Table 7). These data again demonstrate reductions in all classes of measured toxicants, and therefore it is apparent that the TSS is functioning as expected in the EC, to give overall MS toxicant yield reductions.

3.2 Filter Comparisons

From the MS yield data shown m Table 7 all the ECs gave lower yields of carbonyls and vapour phase constituents than the respective commercial comparator cigarettes, with the exception of formaldehyde and 1,3-butadiene yields for BT1. To understand better the contribution of the blend and the selective filters used in the ECs to the overall reductions m these smoke constituents, direct comparisons were made between the ECs and control cigarettes (SC-BT1, SC-TSS1 and SC-TSS6), which were identical in all aspects to the appropriate EC, except for the use of a mono-stage CA filter without adsorbents. The comparisons of the yields from EC and control cigarettes for the carbonyls and other vapour phase constituents are shown in Tables 8 and 9 (shown in Figures 6 and 7, respectively).

From these data it is clear that the yields of the carbonyls and the other vapour phase constituents were all reduced by the presence of the triple stage filter containing CR20L and high activity carbon used in ECs BT1 and TSSl (Table 8). The mean change m MS yield across all volatile constituents measured from BT1 was a reduction of 50% compared to the control cigarette SC-BT1, with a range of 23% reduction for acetaldehyde to 79% reduction for crotonaldehyde. Very similar reductions were obtained with TSSl, which gave a mean reduction of 50%, with a range from 10% reduction m formaldehyde yield to 79% reduction for

crotonaldehyde yield m comparison to SC-TSS1. From Table 9 it is apparent that the dual filter containing additional polymer derived carbon but without the CR20L resin (as used in TSS6), also reduced the yields of the vapour phase smoke constituents by a mean of 48%, with a range from 11% reduction in acetaldehyde yield to 79% reduction for crotonaldehyde yield. Together, these data confirm that the selective filters used m the ECs removed substantial quantities of volatile smoke constituents from cigarette MS, confirming previous studies with the filter adsorbents. For all of the ECs, the MS yields of both formaldehyde and 1,3-butadiene were lower than measured with the scientific contiol cigarettes. The superior performance of the CR20L resin compared to the high activity carbon at foimaldehyde removal from MS can be seen by the greater reduction m the yield of formaldehyde from a higher starting value (53 g/ cig or 53%) in the BTl / SC-BT1 pair compared to the l

reduction (10%) found with the TSS1 / SC-TSS1 pair. Thus, it is clear that the greater formaldehyde yield seen when comparing BTl with the commercial cigarette CCl (Table 7) must be due to differences in blend between these cigarettes. A similar comparison also confirms that the higher 1,3-butadiene yield from BTl compared to CCl is most likely due to an analytical error in the measurement of 1,3 butadiene with CCl.

3.3 Comparison of EC toxicant yields with those from published cigarette brand data

This paper has focused on a comparison of EC toxicant yields with the yields from two commercial comparator cigarettes. However, to fully establish whether the ECs offer reduced machine yields m comparison to conventional commercial cigarettes it is necessary to compare their yields with those from a wider range of cigarettes. The absolute yield values of the ECs described here can be compared with other published data obtained under HCI smoking conditions, namely: (1) (Health Canada (2004) Constituents and emissions reported for cigarettes sold m Canada

http:/ /www.hc-sc.gc.ca/hc~ps/ lt_formats/hecs-sesc/pdf/tobactabac/legislation/ reg/mdust/constitu-eng.pdf (accessed November 2010); data received on request from TRR_RRRT@hc-sc.gc.ca; (2) Counts, M.E. et al. (2005) Smoke composition and predicting relationships for international commercial cigaiettes smoked with three machine-smoking conditions. Regul. Toxicol. Pharmacol. 41, 185-227; and (3); Department of Health Australia and Ageing: http:/ /www.health.gov.au/mternet/ mam/publishing.nsf/Content/tobacco-emis, (accessed, November 2010). It should, however, be noted that such comparisons must be treated with caution due to the known difficulties based on hmited standardisation between laboratories for the analysis of smoke constituents other than NFDPM, nicotine and CO.

The three data sources above were compiled into one dataset to provide a reference set of global cigarette yield data with which to compare the toxicant yields from the ECs described in this study. The full dataset was truncated as follows: first, arsenic, methyl ethyl ketone, nickel and selenium yields were removed from the dataset because yields were not provided by all three sources; second, a number of brands were removed from the dataset due to incomplete, duphcated or erroneous data (two brands m the HC dataset appear to have erroneous (exchanged) toluene and styrene yields; tar, nicotine and CO yields were not provided in the HC dataset for Gitanes KS, and multiple instances of the same yield data were observed m the HC dataset). Finally, reference products were removed from the dataset to ensure that only commercial brands were included. This resulted m a dataset of 120 cigarette brands covering 16 countries or regions. While extensive it is unlikely that this dataset is fully representative of the range of cigarette products on-sale globally, either with respect to the range of design features, or as a representative sample of global brands. However, while it is hmited in these respects, it does constitute a valid comparator set for the toxicant yields for these ECs.

The data was examined to see if it was normally distributed; while a number of toxicants m the dataset were normally distributed the majority (and m particular nitrogenous toxicants such as TSNAs and aromatic amines) were not. Consequently the reference dataset was subject to an empirical cumulative distribution analysis, producing a percentile distribution within the toxicant yields. Yields from the ECs were then compared to the empirical cumulative distribution to identify the position of these yields in comparison to the commercial brands (Figures 8 and 9). In these comparisons, the yields of the ECs described here fall at the low end of the range for numerous toxicants and often give lower values for specific toxicants than any of the products in the commercial bland dataset. Exceptions to this are catechol yields from BTl and NO and TSNA yields from TSSl and TSS6, where the yields are approximately equivalent to the median values for the commercial product dataset. In contrast, the yields of the commercial comparator cigarettes CC1 and CC6 are generally distributed over the range of yields observed with the commercial dataset.

A further comparison was conducted, examining the total toxicant levels from the ECs and each of the commercial products in the dataset. This was conducted in three ways. The first method was to sum the yields of the 39 toxicants for each cigarette to give a total toxicant yield (TTY) for each brand. This approach is of limited utility because the TTY value for each brand is dominated by tar, CO and nicotine, and many other toxicants do not contribute significantly to the total value. A second approach was to sum the yields of all toxicants (but excluding tar, nicotine and CO yields) for each cigarette to give a total for the toxicant subset of yields (TSY). A third, normalisation method gave greater insight into the contribution of all toxicants, wherein a median value was calculated for each toxicant m the commercial dataset. The median value was normalised to 100 for each toxicant, and the yields of toxicants scaled against this value of 100. Totalling the scaled values for all toxicants gave a normalised toxicant total (NTT) for each brand. The TTY, TSY and NTT values for the ECs are compared to and ranked against the values for all of the brands m the commercial dataset m Figures 10 to 12. The comparisons show, with each of the approaches, that the ECs were at the low end of the ranking order. The lmg ECs were found to have the lowest total toxicant yields under each of the three approaches, and the 6mg EC was also lower than any of the commercial brands for the TSY and NTT. In the TTY analysis two of the 120 commercial products have lower TTY values than TSS6 due to their lower tar and nicotine values. The commercial comparator cigarettes CC1 and CC6 were also reasonably low in total toxicant values in comparison with the dataset of commercial brands, falling around the lower quartile of values.

Together these analyses show that the ECs offer some of the lowest machine toxicant yields of cigarettes for which pubhshed HCI smoke chemistry is available; these comparisons therefore confirm that the ECs generate reduced machine toxicant yields m comparison to known levels of commercial cigarettes. 3.4 Comparisons of EC yields as a tatio to nicotine yields

The analysis described above is restiicted to assessment of machine yields of toxicants. However, it has been proposed that the ratio of smoke toxicants to the MS nicotine yield of cigarettes gives a better predictor of smokers' exposure to the toxicant than the MS yield value alone. Therefore, the ratio of MS constituents yields measured m this study to the MS nicotine yields, all measured under HCI smoking machine conditions, has been calculated and is given as a supplemental table (Table 10, Figures 8A and 8B). Under Health Canada Intense machine smoking conditions, the NFDPM yields from BTl, TSS1 and CC1 were comparable, but the nicotine yield from BTl was slightly higher and the nicotine yield from TSS1 shghtly lower than from CC1 (Tables 4 and 7). When the yield values for the EC were calculated as a ratio to the nicotine yield, and compared to those from CC1 and CC6, they followed the same trends as found when comparing the yields per cigarette, but the lower values from BTl when compared to CC1 are more pronounced and the lower values from TSS1 when compared to CC1 are slightly less pronounced.

3.5 Sidestream Smoke Yields

To complete the chemical analysis of smoke emissions from the EC, SS yields for the expanded list of smoke constituents were measured, under ISO smoking parameters. The ISO smoking parameters were chosen because they generate higher SS yields than any of the other smoking regimes. In general, under any smoking regime, the quantity of sidestream smoke can be expected to be dependent on the amount of tobacco consumed in the static burn or smoulder phase of cigarette smoking. The SS yield results are presented as a comparison between the ECs BTl and TSS1 and the commercial cigarette CC1, in Table 10.

Statistically significantly higher yields of sidestream NFDPM (21%), and several constituents such as benzo(a)pyrene (28%), phenohcs (28-77%), carbonyls (22-63%) and volatile hydrocarbon (20-24%) constituents were found with BTl than from CC1. In contrast lower yields of nitrogenous SS smoke constituents such as TSNAs (31-82%), HCN (47%), aromatic amines (21-40%) nitrogen oxides, pyridine and quinolene (19-35%) were found with BTl than with CC1. Most of these changes were described previously (Liu et al, 2010), however, the higher SS phenolic yields and lower than anticipated TSNA yields fiom BTl suggest that chemical differences between Virginia and US-blended tobaccos also influence the SS yields of individual constituents. Finally, the 13% higher tobacco weight from BTl than from CCl will also contribute across the board to the observed increases.

Many SS smoke constituent yields were lower from the EC cigarette TSSl than from CCl . The greatest numeiical differences m SS yields were observed for the TSNAs which were 28 to 52% lower from TSSl than CCl; these observations are consistent with the observed trends m MS yields of these species. The wide range of reductions most likely reflects the reduction m tobacco mass in the cigarettes resulting from incorporation of the TSS, and consequent decrease m the total amount of smoke generated. The one constituent with a statistically significantly higher sidestream yield from TSSl than from CCl was formaldehyde (19% higher). Higher SS foimaldehyde yields were also observed with higher levels of TSS inclusion m the blend (McAdam et al, 2010), suggesting that formaldehyde might be a combustion by-product of the organic materials used m TSS manufacture.

4. Conclusions

Three ECs were made using a combination of technological approaches, and chemical testing under four different machine smoking parameters has confirmed overall reductions of MS toxicants yields from the ECs. When compared with published values of MS toxicant yields from conventional cigarettes, despite elevated formaldehyde yields with BTl, the performance of these ECs appears to be superior, even if they are ranked on a nicotine ratio basis. The data presented in this study support a designation of these ECs as reduced machine-yield prototypes, and previous data with EC made using the TSS approach suggest that lowei biomarkers of exposure to MS toxicants could be achieved with these RMYPs when used by smokers.

Despite the low overall machine yields of toxicants obtained from the cuirent RMYP and their performance against commercial comparators and other published toxicant yield data, substantial amounts of scientific data would need to be acquired, including biomarkers of exposure and biomarkers of biological effect, to determine whether such products might be associated with lower health risks, and therefore there is no certainty that these RMYP will meet the IOM definitions of a PREP.

Nonetheless, we believe that the results from this study are sufficient to encourage further work, including human biomarker studies of these RMYP and further application and refinement of the technologies used in their manufacture.

5. Prototype Smoking Articles

Three prototype RTP smoking articles were produced according to the present invention. The cigarettes are of king size format with a filter length of 27mm and a tobacco rod of 56mm. The prototypes have a tobacco rod comprising a mix of lamina, Expanded Tobacco and non tobacco sheet or modified tobacco.

Conventional cigarette paper is used to form the tobacco rod and ensure the achievement of burn rate and subsequent puff number.

The filter for two of the prototypes is a triple filter composed of a CA mouth end segment (7 mm m length), a CA central segment containing CR20 HD ion exchange resin (10 mm in length) and a dalmation style tobacco end segment containing carbon beads with an engineered microstructure (10 mm m length). The filter for the third prototype is a dual filter composed of a CA mouth end segment (15mm in length) and a dalmation style tobacco end segment containing high activity, polymer-derived carbon beads (12 mm in length). The prototype cigarettes were manufactured to give ISO NFDPM yields of 1 (T562 and H671) and 6 mg (F752). The specification of the prototype cigarettes is described m more detail m Tables 11 to 13.

Table 11. Tobacco blend specifications

Prototype T562 (lmg) H671 (lmg) F752 (6mg)

Lamina (%wwb) 40 12.5 55.0

Expanded Tobacco" 40 12.5

(%wwb)

Expanded Tobacco^b 25.0

(%wwb)

Modified Tobacco^c (%wwb) - 75 - Non tobacco sheet^d (% wb) 20 - 20

Added Top Flavour 0.8 0.8 0.8

(A OLSA) (%wwb)

a Aurora - 100% flue cured tobacco

bSCB - 50% flue cured, 50% Butley

cTobacco processed using the tobacco blend treatment

dThe non tobacco sheet is TSS with the following specification: Chalk (78.5%), Kelvis Alginate (7.5%), Glycerol (12.5%) and Caramel colourant (1.5%)

(manufacturer; De -HTL).

Table 12. Cigarette specifications

Carbon monoxide (mg) 1.0 1.0 4.9

Table 13. Filter specification

Code T562 (lmg) H671 (lmg) F752 (6mg)

Filter Code (Filtrona, USA) SAM 013108- SAM 013108- SAM 020608- 031 031 040

a CR20 HD = amine functionalised resin (manufacturer: Mitsubishi)

b Blucher carbon = spherical carbon beads (manufacturer: Adsor Tech.)

TTugwrap for completed Dual or Triple filter 6. Smoke Toxicant Exposure Study

This study looked at the evaluation of biomarkers of exposure (BoE) to toxicants m smokers who switched from conventional cigarettes to reduced toxicant prototype (RTP) cigarettes according to the present invention.

The technologies discussed in detail above were combined to produce one 6mg and two lmg ISO tar yield RTPs as detailed in Table 14 below.

Table 14: Tested prototype products

Product identifier & Tobacco Tobacco ISO* tar yield HCI# tar description blend filter target yield

(actual) (actual)

CC6 100% US Single 6mg (5.0mg) 24.4mg

Control based on 6mg style tobacco segment:

ISO conventional blend CA

cigarette

TSS6 80% US style Dual 6mg (5.3mg) 20.7mg

6mg ISO tar tobacco segment:

prototype blend CA +

20% tobacco Carbon

substitute (80mg)

sheet CC1 100% US Single lmg (1.2mg) 18.9mg

Control based on lrng style tobacco segment:

ISO tar commercial blend CA

cigaiette

TSS1 80% US style Triple lmg (l .Omg) 17.3mg lmg ISO tar tobacco segment:

prototy e blend CA +

20% tobacco Carbon

substitute (60mg) +

sheet Resin (20mg)

BT1 25% Virginia Tiiple lmg (1.2mg) 17.8mg lmg ISO tar style tobacco segment:

prototype with blend CA +

tobacco blend (untreated) Carbon

modification 75% Virginia (60mg) +

style tobacco Resin (20mg)

blend

(treated)

* ISO regime = 35mL puff volume, 2 second duration, 60 second interval, filter ventilation 100% open

# HCI (Health Canada Intense) regime = 55mL puff volume, 2 second duration, 30 second interval, filter ventilation 100% blocked Smoke chemistry indicated good reductions in toxicants compared to control cigaiettes of conventional design, see Table 15 (Figure 15).

A six week single-centre, single -blinded, randomised controlled switching study with occasional clinical confinement, as illustrated m Figure 16, was conducte. A total of 301 healthy adult subjects were recruited into the study; 100 smokers of 6-7 mg ISO tar yield cigarettes (assigned to the 6 mg groups), 151 smokers of 1-2 mg ISO tar yield cigarettes (assigned to the 1 mg groups) and 50 non-smokers. Recruited smokers were randomly assigned to a control or test group within their tar band, with approximately 50 per group. All smokers smoked a supplied control product for 2 weeks after which Day 14 basehne measurements were made. Control group smokers continued to smoke the control product for a further 4 weeks, while test group smokers were switched to an RTP for 4 weeks. In each case, measurements were made at Days 28 (two weeks) and 41 (four weeks). The non-smoker group provided an indication of background levels of biomarkers. Collection of 24 hour urine samples occurred during three (for smokers) and two (for non-smokers) short periods of chnical confinement (see Figure 16), and exposure to a number of smoke constituents was estimated by analysis of levels of urinary biomarkers of exposure. Analysis of biomarkers of exposure was achieved using vahdated LC-MS/MS methods.

When the RTP smoke chemistry was compared to that of the control cigarette, most measured toxicants were substantially lower (10-96%) with actual levels dependant on design and toxicant (see Table 15). The only higher yields were for one product (BT1) which delivered 16% more nicotine and 35% more 1,3-butadiene, although this was also the product that showed the greatest overall reductions for all other toxicants. Direction and relative magnitude of changes m corresponding biomarkers were largely in-keeping with the changes m smoke chemistry (Table 15 and Figures 17 and 18) although in a few cases a reduction in the smoke was accompanied by an increase in the biomarker (nicotine and NNK for TSS1) or an increase in the smoke but a reduction in biomarker (1,3-butadiene for BT1). Reasons for these

discrepancies are unknown, but may involve analytical variability or smoker behavioui.

Figure 17 shows the biomarker results for Group 2 who switched from control cigarette CC6 (Day 14) to test cigarette TSS6 (Day 41). * denotes a statistically significant difference (p≤ 0.01) between day 14 and 41 results. Non-smoker biomarker levels are shown for reference. All non-smoker levels were significantly lower than day 14 values

Figure 18 shows the biomarker results for Group 4, who switched from control cigarette CC1 (Day 14) to test cigarette TSS1 (Day 41) and Group 5, who switched from control cigarette CC1 (Day 14) to test cigarette BT1 (Day 41).

The study found that, on average, groups of cigarette smokers who switched to reduced toxicant prototype cigarettes had reduced levels in the corresponding biomarker of exposure (BoE). These included BoEs for particulate and vapour phase toxicants. Different prototypes resulted m different levels of reductions to the BoE, in some cases with reductions substantially greater than 50%, depending upon which combination of technologies was used. Generally most of the reduction ιη biomarker level was apparent two weeks after switching. In all cases the average biomarker level was lower m the non-smoker group

This study demonstrates for the first time significant reductions in a range of BoE of tobacco smoke toxicants m smokers following a switch from conventional cigarettes to reduced toxicant prototype cigarettes according to the present invention.

Figure 1 shows a smoking article design according to an embodiment of the present invention. The smoking article 1 comprises a tobacco rod 2 and a filter 3. The tobacco rod comprises a rod of smokeable material, the composition of which is 75% blend treated tobacco, 12.5% leaf and 12.5% expanded tobacco.

The blend treated tobacco is a tobacco with reduced protein and polyphenol content which results from the following process: (l) aqueous extiaction of a tobacco; (u) passing the aqueous extract through a clay and a resin; (iii) treatment of the fibre with an enzyme and deactivation; and (iv) recombming the extract and fibre and drying. The leaf is tobacco as is used in conventional commercial cigaiettes. The expanded tobacco is a tobacco that has been expanded using a supeicritical C0₂ process which is used in conventional commercial cigarettes.

The filter 3 is attached to the tobacco rod 2 by a tipping paper which is a non- porous paper.

The filter 3 is made up of three sections, as indicated by the inset. The section 4 adjacent the end of the tobacco rod is 10 mm in length and contains 60 mg of synthetic carbon. This is a form of carbon which has an engineered porous structure. The middle section 5 is 10 mm in length and contains 20 mg (that is, 2 mg/mm) of CR20HD, an amine functionalized resin having a water content of 12- 17%. The mouth-end section 6 of the filter is 7 mm in length. This may comprise, for example, cellulose acetate tow as used m conventional commercial cigarettes. In possible variations of the smoking article design shown in Figure 19, the smokeable material may further include tobacco substitute sheet. Tobacco substitute sheet is a chalk-based sheet containing glycerol that reduces the quantity of tobacco in a cigarette when incorporated into the tobacco blend. The tobacco substitute sheet may replace some or all of any or all of the different materials making up the smokeable material of the smoking article design discussed above.

A further variation may be to use CR20D in the filter. CR20D is an amine functionalized resin having a water content of 0-5%. For example, CR20D may partially or completely replace the CR20HD used in the design discussed above.

The foregoing description and examples have been set forth merely to illustrate the invention and are not intended to be limiting. Since modifications of the described embodiments incorporating the spirit and substance of the invention may occur to persons skilled in the art, the invention should be construed broadly to include all variations within the scope of the appended claims and equivalents thereof.

Claims

Claims

1. A smoking article comprising at least two of:

(a) a tobacco blend comprising one or more tobaccos or tobacco grades with low TSNA and/ or metal content;

(b) a tobacco blend that has been treated to remove polyphenols and/or peptides;

(c) a tobacco substitute sheet comprising a non-combustible inorganic filler, a binder and an aerosol generating means;

(d) a high activity carbon comprising a polymer-derived carbon material; and

(e) an amine-functionalised chelating resin.

2. A smoking article as claimed in claim 1, comprising a rod of smokeable material comprising up to 60% tobacco substitute sheet.

3. A smoking article as claimed in claim 2, comprising a rod of smokeable material comprising 20% tobacco substitute sheet and 80% tobacco.

4. A smoking article as claimed in any one of the preceding claims, wherein glycerol is subsequently added to blend of the tobacco substitute sheet and tobacco.

4. A smoking article as claimed in claim 1, comprising a rod of smokeable material comprising up to 100% treated tobacco blend.

5. A smoking article as claimed in 4, comprising a rod of smokeable material comprising 75% treated tobacco blend and 25% tobacco.

6. A smoking article as claimed in any one of the preceding claims, comprising a filter having three sections, the mouth end stage comprising cellulose acetate, the middle stage comprising cellulose acetate and an amine-functionalised chelating resin, and the tobacco end stage comprising cellulose acetate and high activity carbon.

7. A smoking article as claimed in claim 6, wherein the middle stage comprises 20 mg of an amine-functionalised chelating resin.

8. A smoking article as claimed in claim 6 or 7, wherein the tobacco end stage comprises 60 mg carbon.

9. A smoking article as claimed m any one of claims 1 to 3, comprising a filter having two sections, the mouth end stage comprising cellulose acetate, and the tobacco end stage comprising cellulose acetate and carbon.

10. A smoking article as claimed in claim 9, wherein the tobacco end stage comprises 80 mg carbon.

11. A smoking article as claimed in any one of claims 6 to 9, wherein the carbon is high activity carbon beads comprising a polymer-derived carbon material.

12. A smoking article as claimed in any one of claims 6 to 9, wherein the amine- functionalised chelating resin is CR20.

13. A smoking article as claimed m claim 1, comprising at least one selected from (a), (b) and (c), and at least one selected from (d) and (e).

14. A smoking article as claimed m claim 13, comprising (c) and (e) .

15. A smoking article as claimed in any one of the preceding claims, having a significant reduction in all mainstream smoke constituents considered to be undesirable.