SUMMARY
Objectives
The public health burden of tobacco use is shifting to the developing world, and the tobacco industry may apply some of its successful marketing tactics, such as allaying health concerns with product modifications. This study used standard smoking machine tests to examine the extent to which the industry is introducing engineering features that reduce tar and nicotine to cigarettes sold in middle- and low-income countries.
Study design
Multicountry observational study.
Methods
Cigarettes from 10 different countries were purchased in 2005 and 2007 with low-, middle- and high-income countries identified using the World Bank’s per-capita gross national income metric. Physical measurements of each brand were tested, and tobacco moisture and weight, paper porosity, filter ventilation and pressure drop were analysed. Tar, nicotine and carbon monoxide emission levels were determined for each brand using International Organization for Standardization and Canadian Intensive methods. Statistical analyses were performed using Statistical Package for the Social Sciences.
Results
Among cigarette brands with filters, more brands were ventilated in high-income countries compared with middle- and low-income countries [χ2(4)=25.92, P<0.001]. Low-income brands differed from high- and middle-income brands in engineering features such as filter density, ventilation and paper porosity, while tobacco weight and density measures separated the middle- and high-income groups. Smoke emissions differed across income groups, but these differences were largely negated when one accounted for design features.
Conclusions
This study showed that as a country’s income level increases, cigarettes become more highly engineered and the emissions levels decrease. In order to reduce the burden of tobacco-related disease and further effective product regulation, health officials must understand cigarette design and function within and between countries.
Keywords: Tobacco, Emissions, Regulation
Introduction
The public health burden of tobacco use is shifting towards the developing world, such that by 2030, more than 80% of the world’s tobacco-related deaths will be in developing countries.1 This change is occurring, in part, because the tobacco manufacturers appear to be applying some of the same marketing tactics that made them successful in recruiting and retaining customers in high-income countries in middle- and low-income countries. One can anticipate that cigarette manufacturers will, as a matter of course, continually update their product lines to adapt to shifting markets and consumer needs.2 As public awareness of the health risks of smoking gradually increased over the past 50 years, tobacco manufacturers responded by introducing filtered and so-called ‘low tar and nicotine’ brands into the marketplace. Trends towards lower tar brands worldwide are likely in response to consumers’ growing awareness of the negative health effects of smoking.3,4 The marketing of lower tar cigarettes is unfortunately aided by well-meaning government regulations that include International Organization for Standardization (ISO)/Federal Trade Commission tar, nicotine and carbon monoxide (TNCO) yields on packs, which merely draws attention to meaningless distinctions among brands.5 Labelling packages with this information is troublesome because TNCO emission numbers are misleading to consumers, who do not realize that these numbers are not related to the level of risk for a particular product.
Cigarette design plays a major role in determining TNCO yields in tests using machine smoking. Cigarette filters, usually made from cellulose acetate filaments, are able to reduce the machine yields of tar and nicotine by 40–50%.6 In high-income countries, such as the USA, filtered cigarettes became popular in the early 1950s as the demand for lower smoke yields grew, following the initial reports of lung cancer associated with smoking, and now represent the majority of the market.6 The addition of filter ventilation to cigarettes is a method widely used by cigarette manufacturers to achieve lower machine-measured TNCO yields, able to achieve lower smoke yields than filters, paper permeability and tobacco processing would be able to achieve alone.6 Ventilation holes appear as a ring of tiny perforations that circle the filter end of the cigarette. This design feature allows air to enter the holes when a puff is taken, which reduces the amount of tobacco which is consumed, and dilutes the tobacco smoke coming from the mouth end.6 Filter ventilation is successful in lowering TNCO yields when cigarettes are smoked by machines for testing, but they are not as successful for people. People knowingly and unknowingly subvert this design feature by blocking the ring of holes by a number of different methods. The holes are blocked by smokers’ fingers, lips and, in some cases, even tape.7 Smokers may also defeat ventilation by taking larger, more frequent puffs or smoking to a shorter butt length to compensate.7 Internal documents available as a result of litigation reveal that the tobacco industry capitalized on the limitations of the smoking machine protocol used to report TNCO yields to consumers and regulators; cigarettes were promoted as ‘low yield’ when they only produced low yield under unrealistic smoking conditions.8
Given the growth of tobacco use in the developing world and the history of tobacco product design in developed countries, the purpose of the current study is to examine whether the design of cigarettes varies by country development level. It was hypothesized that countries in different stages of the tobacco epidemic and awareness of the health risks of smoking would have cigarettes that differ in their design features.
Methods
Cigarette acquisition and storage
The cigarettes analysed in this project originated in 10 different countries (Table 1). Prior research had focused almost exclusively on cigarettes from the USA, Canada, the UK and Australia,9,10 so cigarettes from these markets were not considered in the current study. Field teams in each country purchased 18 packs each of popular brands of cigarettes in that market from three distinct retail locations. Packs were purchased in the Czech Republic and Greece in 2005, and packs from the remaining countries were purchased in 2007. The packs were shipped via common courier to Roswell Park Cancer Institute, where they were catalogued and stored at −20°C until analysis. In accordance with ISO 3402:1999, the packs were conditioned for a minimum of 48 h at 22±2.0°C and 60±2.0% relative humidity in an environmental chamber prior to testing.
Table 1.
Summary of countries, income levels and brands studied
| Country | Income level | Number of brands | FCTC Status | Leading manufacturera | Emissions, content and design regulations |
|---|---|---|---|---|---|
| New Zealand | High | 11 | Party | BAT | None |
| Cyprus | High | 12 | Party | BAT | Subject to EU 10:1:10 limits on tar, nicotine and CO emissions; yields displayed on packs |
| Greece | High | 10 | Party | Philip Morris | Subject to EU 10:1:10 limits on tar, nicotine and CO emissions; yields displayed on packs |
| Czech Republic | High | 9 | Signatory | Philip Morris | Subject to EU 10:1:10 limits on tar, nicotine and CO emissions; yields displayed on packs |
| Mexico | Middle | 15 | Party | Philip Morris | Yields displayed on packsb |
| Turkey | Middle | 15 | Party | Philip Morris | Yields displayed on packs |
| Romania | Middle | 10 | Party | BAT | Subject to EU 10:1:10 limits on tar, nicotine and CO emissions; yields displayed on packs |
| Nepal | Low | 15 | Party | Local | None |
| Indonesia | Low | 15 | Not signed or ratified | Philip Morris | Yields displayed on packs |
| India | Low | 6 | Party | Local | Authority exists to set maximum limits and require reporting |
| Total | 118 |
FCTC, Framework Convention on Tobacco Control; BAT, British-American Tobacco; EU, European Union; CO, carbon monoxide.
Data from Shafley O, Eriksen M, Ross H, Mackay J. Tobacco atlas. 3rd edn. Atlanta: American Cancer Society; 2009.
At time of study. Per regulations introduced in 2009, yields are no longer displayed on packs.
Income categories
The World Bank divides national economies according to gross national income (GNI) per capita, expressed in US dollarsa. The countries chosen for comparison were divided into low-income (up to $3595), middle-income ($3596–11,115) and high-income ($11,116 or more) categories on this basis.
Physical measurements
Five cigarettes were selected from each pack after conditioning for physical analysis. Digital calipers were used to measure the length of the entire cigarette, the length and diameter of the tobacco rod, and the length and diameter of the filter. Filter weight measurements were made gravimetrically using an analytical balance. The tipping paper was removed from each filter, and measurements of its length and the presence of any vent holes were performed using a transparent ruler and a light box.
Tobacco moisture and tobacco weight
Tobacco moisture and tobacco weight were analysed using a HR83 Moisture Analyzer (Mettler-Toledo, Columbus, OH, USA), and procedures were adapted from the Mettler-Toledo Application Data Sheetb. The moisture content was determined as the percentage change in weight after heating the tobacco from five cigarettes with a halogen bulb at 125°C.
Paper porosity
The level of porosity of the cigarette paper was measured using a PPM1000M paper porosity device (Cerulean, Milton Keynes, UK). Five cigarette papers were tested for each brand, with five readings taken for each paper, and the 25 total measurements were averaged.
Filter ventilation and pressure drop
Measurements of the cigarette filter ventilation and pressure drop were taken using a KC3 combined dilution/pressure drop instrument (Borgwaldt-KC, Richmond, VA, USA). The average was taken from five cigarettes tested with this device.
Tar, nicotine and CO emissions
The determination of TNCO emissions levels was performed by an independent laboratory (Labstat, Kitchener, ON, Canada) using both ISO and Canadian Intensive (CI) methods. The international standard method, ISO 3308, requires a puff volume of 35 ml, a puff duration of 2.0 s and a puff frequency of one per 60 s. Butt length is specified as 23 mm for non-filter cigarettes and the length of filter overwrap plus 3 mm for filtered cigarettes. A modified protocol is used to test the cigarette under ‘intense’ smoking conditions. According to Health Canada’s intense smoking protocol, 100% of the filter vents are blocked by a single layer of tape covering the filter and tipping paper. The puff volume is 55 ml and the puff frequency is 30 s, but other parameters are identical to ISO 3308.
Statistical analysis
Data analysis was completed using Statistical Package for the Social Sciences Version 14.0 (SPSS Inc., Chicago, IL, USA). Pearson correlations were used to identify univariate associations between physical and design measurements and TNCO emissions under both the ISO and CI regimens, and for high-, middle- and low-income countries. Discriminant function analysis was used to determine design features that best discriminated among the three development levels, with Wilks’ lambda used to test for significance. Multivariate repeated-measures analysis of variance was used to examine associations between development level and TNCO emissions, including adjustment for design features. Here, Hotelling’s trace F-test for multivariate significance was used, with the Greenhouse–Geisser correction for violations of the sphericity assumption applied in univariate tests.
Results
Table 2 shows the mean values for design features on cigarettes across countries. Initial analysis revealed that all of the brands from high-income countries had filters, compared with 95% of the middle-income brands and 86% of the low-income brands with filters. If one breaks this out further into whether those filters are ventilated, a pattern emerges where 95% of cigarettes in high-income countries had vented filters, compared with 87.5% in middle-income countries and 44.4% in low-income countries; this difference was statistically significant [χ2(4)=25.92, P<0.001]. The majority of products popular in Indonesia were kreteks, which differed significantly in terms of several characteristics (higher tobacco rod length, weight, density and filter density, all P<0.001) from other products purchased in the other low-income countries. For subsequent multivariate analyses, only filtered, standard diameter cigarettes were considered. This reduced the overall sample size to 99 varieties.
Table 2.
Basic differences in physical parameters by income level
| Mean | SE | Minimum | Maximum | ANOVA | P | ||
|---|---|---|---|---|---|---|---|
| Cigarette length (mm) | Low | 81.02 | 1.27 | 68.33 | 90.30 | F(2,115)=4.060 | 0.020 |
| Middle | 85.20 | 1.32 | 68.86 | 99.72 | |||
| High | 84.52 | 0.58 | 79.36 | 98.61 | |||
| Rod diameter (mm) | Low | 7.75 | 0.09 | 6.71 | 9.93 | F(2,115)=3.747 | 0.027 |
| Middle | 7.92 | 0.04 | 7.57 | 9.92 | |||
| High | 7.64 | 0.08 | 5.41 | 8.07 | |||
| Tobacco length (mm) | Low | 64.76 | 1.00 | 56.90 | 81.69 | F(2,115)=4.515 | 0.013 |
| Middle | 62.94 | 0.98 | 54.93 | 73.20 | |||
| High | 61.12 | 0.55 | 56.02 | 70.13 | |||
| Length of tipping paper (mm) | Low | 25.61 | 0.80 | 17.56 | 32.31 | F(2,108) =6.109 | 0.003 |
| Middle | 28.54 | 0.69 | 18.09 | 35.86 | |||
| High | 28.60 | 0.53 | 20.21 | 34.78 | |||
| Filter length (mm) | Low | 18.98 | 0.79 | 11.06 | 27.10 | F(2,103)=5.434 | 0.006 |
| middle | 21.38 | 0.87 | 9.22 | 27.12 | |||
| High | 22.61 | 0.66 | 6.85 | 30.21 | |||
| Filter weight (g) | Low | 0.10 | 0.01 | 0.05 | 0.16 | F(2,102)=6.791 | 0.002 |
| middle | 0.12 | 0.01 | 0.05 | 0.16 | |||
| High | 0.12 | 0.00 | 0.04 | 0.16 | |||
| Number of vent rows | Low | 0.75 | 0.23 | 0.00 | 6.00 | F(2,6781)=8.0212.966 | 0.00 |
| Middle | 1.26 | 0.22 | 0.00 | 4.00 | |||
| High | 2.64 | 0.56 | 0.00 | 6.00 | |||
| Distance (mm) – mouth end to nearest row of vents | Low | 13.15 | 0.47 | 11.00 | 15.70 | F(2,55)=4.928 | 0.011 |
| Middle | 12.09 | 0.24 | 10.30 | 14.55 | |||
| High | 11.73 | 0.25 | 10.20 | 15.00 | |||
| Per-cigarette tobacco weight (g) | Low | 0.84 | 0.05 | 0.64 | 1.86 | F(2,115)=12.71 | 0.00 |
| Middle | 0.69 | 0.01 | 0.59 | 0.90 | |||
| High | 0.65 | 0.01 | 0.37 | 0.80 | |||
| Ventilation (%) | Low | 11.52 | 2.51 | 0.00 | 47.04 | F(2,99)=19.188 | 0.00 |
| Middle | 25.95 | 1.68 | 1.30 | 40.70 | |||
| High | 36.94 | 3.46 | 1.56 | 81.94 | |||
| Paper porosity (CORESTA) | Low | 36.40 | 2.07 | 11.42 | 60.21 | χ2(3)=11.07 | 0.004a |
| Middle | 56.87 | 7.88 | 22.70 | 334.31 | |||
| High | 45.17 | 2.48 | 21.05 | 72.45 |
SE, standard error; ANOVA, analysis of variance.
By Kuskal-Wallis test.
Multivariate analysis
To determine whether some combination of design features could be used to distinguish between filtered cigarettes sold in countries of differing income levels, a discriminant function analysis was conducted. Two discriminant functions were derived, with a Wilks’ Lambda value of 0.275 [χ2(24)=108.99, P<0.001]. A significant association between putative groups and predictors was seen [λ=0.615, χ2(11)=41.145, P<0.001]. The first discriminant function (comprised of filter density, ventilation, moisture, filter weight, paper porosity, tipping paper length, rod density, filter diameter, filter length) maximally separates the low-income products from the remaining groups, while the second function (rod diameter, tobacco length, tobacco weight) separates middle- from high-income products, with low-income products falling in between. The two functions accounted for 66.3% and 33.7% of the variance, respectively. Classification statistics indicated an overall accuracy of 78.5%, with 71.4% of low-income, 80.6% of middle-income and 80.6% of high-income cases correctly classified.
A second approach examined whether brands made by the three largest multinational producers (Philip Morris, British-American Tobacco, Japan Tobacco International; n=57), available in countries at all income levels, were systematically different from brands produced by other manufacturers (n=36). A two-way multivariate analysis of variance examined the effects of income level and manufacturer on the set of cigarette design variables. Overall, a significant multivariate interaction effect of manufacturer*income level was observed [F(24,150)=3.428, P<0.001, partial η2=0.354], along with main effects of both manufacturer [F(12,76)=4.754, P<0.001, partial η2=0.429] and income level [F(24,150)=5.496, P<0.001, partial η2=0.468]. Drilling down to univariate effects, the interaction was statistically significant for ventilation, filter density, filter weight and moisture (Table 3). The brands from the major manufacturers were observed to be, on average, largely identical across countries with different income levels, while the other brands showed greater variation.
Table 3.
Mean values for selected design parameters by income level and manufacturer type
| Dependent variable | Manufacturer | Income level | Mean | SE | F(2,93) | P | η2 |
|---|---|---|---|---|---|---|---|
| Ventilation | Other | Low | 7.724 | 3.562 | 14.803 | <0.001 | 0.254 |
| Middle | 23.238 | 4.215 | |||||
| High | 49.187 | 3.848 | |||||
| PM/BAT/JTI | Low | 24.751 | 5.038 | ||||
| Middle | 26.945 | 2.614 | |||||
| High | 26.582 | 2.721 | |||||
| Filter density | Other | Low | 97.648 | 2.219 | 10.869 | <0.001 | 0.200 |
| Middle | 115.402 | 2.625 | |||||
| High | 121.274 | 2.397 | |||||
| PM/BAT/JTI | Low | 112.361 | 3.138 | ||||
| Middle | 113.674 | 1.628 | |||||
| High | 113.628 | 1.695 | |||||
| Filter weight (g) | Other | Low | 0.078 | 0.007 | 4.937 | 0.009 | 0.102 |
| Middle | 0.119 | 0.009 | |||||
| High | 0.117 | 0.008 | |||||
| PM/BAT/JTI | Low | 0.126 | 0.010 | ||||
| Middle | 0.117 | 0.005 | |||||
| High | 0.126 | 0.006 | |||||
| % Moisture | Other | Low | 14.461 | 0.242 | 4.465 | 0.014 | 0.093 |
| Middle | 14.718 | 0.286 | |||||
| High | 15.780 | 0.261 | |||||
| PM/BAT/JTI | Low | 15.194 | 0.342 | ||||
| Middle | 16.144 | 0.177 | |||||
| High | 15.819 | 0.185 | |||||
PM, Philip Morris; BAT, British-American Tobacco; JTI, Japan Tobacco International; SE, standard error.
Examination of tar, nicotine and CO yields
Emissions for all brands were compared under the ISO and CI smoking regimens; results of these basic comparisons are shown in Table 4. In general, cigarettes in the low-income countries were found to have significantly higher puff counts compared with products from middle- and high-income markets regardless of regimen. To account for this difference, yields were indexed per litre of smoke for subsequent analysis, following a recommendation of Rickert et al.11
Table 4.
Differences in International Organization for Standardization (ISO) and Health Canada puff counts and yields by country income level
| Mean | SE | Minimum | Maximum | ANOVA | P | ||
|---|---|---|---|---|---|---|---|
| ISO Puff count | Low | 10.12 | 0.64 | 6.49 | 21.53 | F(2,110)=13.17 | 0.00 |
| Middle | 7.70 | 0.21 | 5.05 | 10.76 | |||
| High | 7.75 | 0.11 | 6.31 | 9.90 | |||
| ISO Tar | Low | 22.27 | 2.18 | 6.56 | 60.03 | F(2,110)=36.79 | 0.00 |
| Middle | 10.69 | 0.25 | 8.04 | 13.78 | |||
| High | 7.62 | 0.62 | 0.46 | 13.82 | |||
| ISO Nicotine | Low | 1.33 | 0.09 | 0.62 | 2.81 | F(2,110)=33.18 | 0.00 |
| Middle | 0.85 | 0.03 | 0.45 | 1.16 | |||
| High | 0.81 | 0.07 | 0.12 | 1.18 | |||
| ISO CO | Low | 13.23 | 0.70 | 7.05 | 23.21 | F(2,91)=20.33 | 0.00 |
| Middle | 9.44 | 0.27 | 5.16 | 12.38 | |||
| High | 7.15 | 0.54 | 0.64 | 12.90 | |||
| Canadian puff count | Low | 13.37 | 0.84 | 8.35 | 27.06 | F(2,110)=18.17 | 0.00 |
| Middle | 9.89 | 0.28 | 7.42 | 14.46 | |||
| High | 9.32 | 0.23 | 6.50 | 13.30 | |||
| Canadian Intense tar | Low | 47.18 | 3.44 | 28.09 | 110.07 | F(2,110)=28.78 | 0.00 |
| Middle | 31.34 | 0.57 | 23.06 | 38.47 | |||
| High | 27.02 | 0.68 | 18.09 | 34.80 | |||
| Canadian Intense nicotine | Low | 2.75 | 0.15 | 1.60 | 5.94 | F(2,110)=17.87 | 0.00 |
| Middle | 2.13 | 0.08 | 0.95 | 2.98 | |||
| High | 1.92 | 0.06 | 1.22 | 2.54 | |||
| Canadian Intense CO | Low | 28.47 | 1.24 | 19.57 | 51.41 | F(2,110)=5.83 | 0.00 |
| Middle | 25.32 | 0.56 | 11.96 | 30.74 | |||
| High | 24.99 | 0.39 | 19.09 | 30.20 |
CO, carbon monoxide; SE, standard error; ANOVA, analysis of variance.
A multivariate repeated-measures analysis of variance was used to examine the change in TNCO emissions per litre of smoke simultaneously among the three income levels. (As above, unfiltered cigarettes, kreteks and superslims were omitted from the analysis.) A significant multivariate difference in yields by development level was found [F(6,176)=11.166, P<0.001, partial η2=0.276], along with a main effect of testing method (ISO vs CI) [F(3,89)=136.596, P<0.001, partial η2=0.822] and a significant interaction [F(6,176)=9.286, P<0.001, partial η2=0.240]. Univariate results showed that TNCO yields increased from ISO to CI methods, as expected, but the patterns of change depended on income level to differing extents.
To examine the impact of design features on these relative differences between ISO and CI yields adjusted for smoke volume, the design features were entered into the model as covariates. A backward elimination approach was taken to remove those design features that did not contribute explanatory power from the model, with a criterion of P>0.10 for removal. The final model included tobacco length, tipping paper length, paper porosity, ventilation, tobacco weight and filter density. Even with the inclusion of these covariates, development level remained significantly associated with emissions per litre (Table 5). However, accounting for the cigarette design variables eliminates most of the differences observed between ISO and CI TNCO levels between high-, middle- and low-income countries.
Table 5.
Multivariate between- and within-subject tests for development level and cigarette design parameters with respect to tar, nicotine and carbon monoxide per litre of smoke under International Organization for Standardization (ISO) and Canadian Intense (CI) testing conditions
| Variable | Hotelling’s trace | F | Test df | Error df | Sig. | Partial η2 |
|---|---|---|---|---|---|---|
| Intercept | 2.921 | 75.940 | 3 | 78 | <0.001 | 0.745 |
| Income group | 0.379 | 4.858 | 6 | 154 | <0.001 | 0.159 |
| Ventilation | 1.063 | 27.649 | 3 | 78 | <0.001 | 0.515 |
| Tobacco weight | 0.402 | 10.462 | 3 | 78 | <0.001 | 0.287 |
| Tipping paper length | 0.374 | 9.713 | 3 | 78 | <0.001 | 0.272 |
| Tobacco rod length | 0.258 | 6.715 | 3 | 78 | <0.001 | 0.205 |
| Paper porosity | 0.239 | 6.222 | 3 | 78 | 0.001 | 0.193 |
| Filter density | 0.2 | 5.204 | 3 | 78 | 0.002 | 0.167 |
| Smoking method (CI vs ISO) | 0.231 | 6.017 | 3 | 78 | 0.001 | 0.188 |
| Method * income group | 0.503 | 6.452 | 6 | 154 | <0.001 | 0.201 |
| Method * vent | 9.037 | 234.954 | 3 | 78 | <0.001 | 0.900 |
| Method * tobacco weight | 0.343 | 8.917 | 3 | 78 | <0.001 | 0.255 |
| Method * porosity | 0.272 | 7.062 | 3 | 78 | <0.001 | 0.214 |
| Method * rod length | 0.23 | 5.979 | 3 | 78 | 0.001 | 0.187 |
| Method * tipping | 0.134 | 3.497 | 3 | 78 | 0.019 | 0.119 |
| Method * filter density | 0.107 | 2.772 | 3 | 78 | 0.047 | 0.096 |
As a secondary approach to the question of how design influences emissions, the study examined whether income levels differed in terms of average design features at a given tar yield; the 9–11 mg range was chosen given the European Union upper limit. For this tar level, significant differences were observed by income level in filter density and pressure drop, both of which increased with higher income (data not shown).
Discussion
This study set out to compare the physical design characteristics and smoke emissions of cigarettes from high-, middle- and low-income countries, and found that cigarette design does differ between countries with low, middle and high levels of income. The general trend is that as the country’s income level increases, cigarettes become more highly engineered and the emissions levels decrease. This was confirmed by discriminant analysis, where a cluster of engineering elements best distinguished cigarettes from low-income countries from those in middle- and high-income countries.
Multivariate analysis showed that when TNCO were adjusted for the amount of smoke produced (i.e. expressed per litre), income level differences persisted, although they were minimized by adjustment for a cluster of cigarette engineering features, the key being filter ventilation. This is consistent with results observed in other studies.12–15 Increased filter ventilation decreases emissions levels in ISO machine testing, albeit in an illusory way, and lower emissions levels are desirable in areas under pressure from health concerns (and, in some cases, government regulations) to produce ‘less harmful’ cigarettes.6,16 Even after these adjustments, cigarettes from lower-income countries still emitted more nicotine. This may be a further marker of design, as the cigarettes from lower-income countries, particularly those made by local manufacturers, may be less likely to contain engineered tobaccos such as reconstituted sheet and expanded tobacco, with which rod nicotine content can be altered.17
Clearly, a wide range of cigarette characteristics exists, and the constellations of characteristics appear to differ across country income levels. The design features appear to explain most of the difference in TNCO yields observed. This reinforces the point that cigarettes can be manipulated by manufacturers to conform to local tastes and product regulatory policies. The World Health Organization Framework Convention on Tobacco Control calls for future regulation of the contents and emissions of products under Article 9,8 and a firm scientific basis is needed to support such regulation. It is clear that the requirements of any regulatory scheme for tobacco products needs to encompass product design as well as toxicant emissions. At the Third Conference of Parties (2008), the working group charged with developing guidelines for Articles 9 and 10 noted that design features of products should be included in any disclosure and testing requirements. In particular, they identified product dimensions, tipping paper length, ventilation, paper porosity, moisture content, packing density, and filter length and density as key metric for which reliable measures exist.18 In addition, the TobReg proposalfor upper limits on specific harmful constituents addresses issues of variable cigarette design in part by adopting a machine smoking regimen that occludes vents and using nicotine-normalized emissions.19 However, other researchers have gone further to call for proscriptions on the use of certain design features, primarily filter vents, by manufacturers.16,20,21 These approaches are not necessarily in conflict. Where deliberate efforts are proposed to reduce toxicants, regulators ought to have some say over what methods for compliance are or are not legitimate. Indeed, it is reasonable to question why, in the case of cigarettes, which are exceedingly dangerous with no accompanying societal benefit, manufacturers should be permitted to continue designing their products without restriction.
Given the growing concentration of the cigarette market in a relatively small number of transnational corporations,22 it is likely that engineered cigarettes which provide an illusion of safety will only become more and more common in low- and middle-income countries unless regulators intercede. Indeed, this study found that products made by the major transnationals were practically identical across income levels. Inaction could leave smokers there vulnerable to the same marketing tactics used in Western markets since the 1970s.6,21,23 Indeed, there is evidence that pack colours, for example, can signal product strength to smokers,24,25 even if actual design differences between products are minimal.26
This research may be limited by the sample of cigarette brands used for the study. It would be unreasonable to test every brand of cigarettes available in each country that participated, so the brands selected may not be a perfectly representative sample. Some brands were chosen according to popularity and market share, as determined by the representative sending the brands to the laboratory. This could affect the results if cigarettes with particular characteristics were over- or under-represented in the sample. Also, the current study did not examine other aspects of design, such as the use of additives, which are also believed to impact smokers’ perceptions of products.27
Innovations in product design are motivated by a variety of factors, including increasing health concerns, public health regulations and manufacturing cost. In order to continue making progress in the improvement of public health by reducing the burden of tobacco-related disease, health officials must possess strong scientific understanding of cigarette design and function, and the ways in which products in different markets are similar and divergent. This way, effective, relevant and enforceable regulations may be placed on tobacco products.
Acknowledgments
The authors thank the research teams in each country for procuring the cigarettes used in this research project.
Funding
This work was supported by the US National Cancer Institute via the Roswell Park Cancer Institute Transdisciplinary Tobacco Use Research Center (P50CA111236) and by P01CA138389. The study sponsor had no role in study design, collection, analysis and interpretation of data, writing of the manuscript, nor the decision to submit the manuscript for publication.
Footnotes
Ethical approval
None sought.
Competing interests
KMC has provided expert testimony in court cases involving the tobacco industry.
http://us.mt.com/global/en/home/supportive_content/application_methods/CigaretteTobacco.rxHgAwXLuMvM.MediaFileComponent.html/CigaretteTobacco_en.pdf (last accessed 10/09/2009).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.World Health Organization. WHO Report on the global tobacco epidemic, 2008 – the MPOWER package. [last accessed 16/06/2010]; Available at: http://www.who.int/tobacco/mpower/2008/en/index.html.
- 2.Wayne GF, Connolly GN. Regulatory assessment of brand changes in the commercial tobacco product market. Tob Control. 2009;18:302–9. doi: 10.1136/tc.2009.030502. [DOI] [PubMed] [Google Scholar]
- 3.Euromonitor International. The world market for tobacco. Chicago: Euromonitor International; 2006. [Google Scholar]
- 4.Euromonitor International. Tobacco – world. Chicago: Euromonitor International; 2007. [Google Scholar]
- 5.Hammond D, Wiebel F, Kozlowski LT, Borland R, Cummings KM, O’Connor RJ, McNeill A, Connolly GN, Arnott D, Fong GT. Revising the machine smoking regime for cigarette emissions: implications for tobacco control policy. Tob Control. 2007;16:8–14. doi: 10.1136/tc.2005.015297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kozlowski LT, O’Connor R. Cigarette filter ventilation is a defective design because of misleading taste, bigger puffs, and blocked vents. Tob Control. 2002;11:40–50. doi: 10.1136/tc.11.suppl_1.i40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hammond D, Collishaw NE, Callard C. Secret science: tobacco industry research on smoking behavior and cigarette toxicity. Lancet. 2006;367:781–7. doi: 10.1016/S0140-6736(06)68077-X. [DOI] [PubMed] [Google Scholar]
- 8.World Health Organization. [last accessed 16/06/2010];WHO Framework convention on tobacco control. 2005 Available at: http://www.who.int/tobacco/framework/WHO_FCTC_english.pdf.
- 9.O’Connor RJ, Hammond D, McNeill A, King B, Kozlowski LT, Giovino GA, Cummings KM. How do different cigarette design features influence the standard tar yields of popular cigarette brands sold in different countries? Tob Control. 2008;17(Suppl 1):i1–5. doi: 10.1136/tc.2006.019166. [DOI] [PubMed] [Google Scholar]
- 10.Kozlowski LT, Mehta NY, Sweeney CT, Schwartz SS, Vogler GP, Jarvis MJ, West RJ. Filter ventilation and nicotine content of tobacco in cigarettes from Canada, the United Kingdom, and the United States. Tob Control. 1998;7:369–75. doi: 10.1136/tc.7.4.369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rickert WS, Collishaw NE, Bray DF, Robinson JC. Estimates of maximum or average cigarette tar, nicotine, and carbon monoxide yields can be obtained from yields under standard conditions. Prev Med. 1986;15:82–91. doi: 10.1016/0091-7435(86)90038-1. [DOI] [PubMed] [Google Scholar]
- 12.Counts ME, Hsu FS, Laffoon SW, Dwyer RW, Cox RH. Mainstream smoke constituent yields and predicting relationships from a worldwide market sample of cigarette brands: ISO smoking conditions. Regul Toxicol Pharmacol. 2004;39:111–34. doi: 10.1016/j.yrtph.2003.12.005. [DOI] [PubMed] [Google Scholar]
- 13.Calafat AM, Polzin GM, Saylor J, Richter P, Ashley DL, Watson CH. Determination of tar, nicotine, and carbon monoxide yields in the mainstream smoke of selected interrnational cigarettes. Tob Control. 2004;13:45–51. doi: 10.1136/tc.2003.003673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Stephens WE. Dependence of tar, nicotine and carbon monoxide yields on physical parameters: implications for exposure, emissions control and monitoring. Tob Control. 2007;16:170–6. doi: 10.1136/tc.2006.017491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.O’Connor RJ, Li Q, Stephens WE, Hammond D, Elton-Marshall T, Cummings KM, Giovino G, Fong GT. Cigarettes sold in China: design, emissions, and metals. Tob Control. doi: 10.1136/tc.2009.030163. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kozlowski LT, O’Connor RJ, Giovino GA, Whetzel CA, Pauly J, Cummings KM. Maximum yields might improve public health if filter vents were banned: a lesson from the history of vented filters. Tob Control. 2006;15:262–6. doi: 10.1136/tc.2006.016501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hoffmann D, Hoffmann I. The changing cigarette, 1950–1995. J Toxicol Environ Health. 1997;50:307–64. doi: 10.1080/009841097160393. [DOI] [PubMed] [Google Scholar]
- 18.Elaboration of guidelines for implementation of Articles 9 and 10 of the WHO Framework Convention on Tobacco Control: progress report of the working group. Conference of the Parties to the WHO Framework Convention on Tobacco Control, Third session; Durban, South Africa. 17–22 November 2008; [last accessed 16/06/2010]. Available at: http://apps.who.int/gb/fctc/PDF/cop3/FCTC_COP3_6-en.pdf. [Google Scholar]
- 19.Burns DM, Dybing E, Gray N, Hecht S, Anderson C, Sanner T, O’Connor R, Djordjevic M, Dresler C, Hainaut P, Jarvis M, Opperhuizen A, Straif K. Mandated lowering of toxicants in cigarette smoke: a description of the World Health Organization TobReg proposal. Tob Control. 2008;17:132–41. doi: 10.1136/tc.2007.024158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.King B, Borland R. The ‘low-tar’ strategy and the changing construction of Australian cigarettes. Nicotine Tob Res. 2004;6:85–94. doi: 10.1080/14622200310001656907. [DOI] [PubMed] [Google Scholar]
- 21.Borland R, Fong GT, Yong HH, Cummings KM, Hammond D, King B, Siahpush M, McNeill A, Hastings G, O’Connor RJ, Elton-Marshall T, Zanna MP. What happened to smokers’ beliefs about light cigarettes when ‘light/mild’ brand descriptors were banned in the UK? Findings from the International Tobacco Control (ITC) Four Country Survey. Tob Control. 2008;17:256–62. doi: 10.1136/tc.2007.023812. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.World Health Organization. Tobacco industry interference with tobacco control. Geneva: WHO Press; 2008. [Google Scholar]
- 23.Elton-Marshall T, Fong GT, Zanna MP, Jiang Y, Hammond D, O’Connor RJ, Yong H-H, Li L, King B, Li Q, Borland R, Cummings KM, Driezen P. Beliefs about the relative harm of ‘light’ and ‘low tar’ cigarettes: findings from the International Tobacco Control (ITC) China survey. Tob Control. doi: 10.1136/tc.2008.029025. (in press) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hammond D, Parkinson C. The impact of cigarette package design on perceptions of risk. J Public Health (Oxf) 2009;31:345–53. doi: 10.1093/pubmed/fdp066. [DOI] [PubMed] [Google Scholar]
- 25.Hammond D, Dockrell M, Arnott D, Lee A, McNeill A. Cigarette pack design and perceptions of risk among UK adults and youth. Eur J Public Health. 2009;19:631–7. doi: 10.1093/eurpub/ckp122. [DOI] [PubMed] [Google Scholar]
- 26.King B, Borland R, Abdul-Salaam S, Polzin G, Ashley D, Watson C, O’Connor RJ. Divergence between strength indicators in packaging and cigarette engineering: a case study of Marlboro varieties in Australia and the USA. Tob Control. doi: 10.1136/tc.2009.033217. (in press) [DOI] [PubMed] [Google Scholar]
- 27.Rabinoff M, Caskey N, Rissling A, Park C. Pharmacological and chemical effects of cigarette additives. Am J Public Health. 2007;97:1981–91. doi: 10.2105/AJPH.2005.078014. [DOI] [PMC free article] [PubMed] [Google Scholar]