INTRODUCTION

Despite substantial progress in reducing prevalence, cigarettes remain the most used tobacco product and carry the greatest association with morbidity and mortality¹. Cigarette manufacturers continue to introduce new brand varieties to retain people who currently smoke and attract new users. Specific design features such as filter ventilation contribute to sensory experiences during smoking and may convey perceptions of reduced health risk². Filter ventilated cigarettes, however, do not promote risk reduction, as their use elicits compensatory behavior, including taking larger puffs and blocking vents². Additionally, in the US, findings from the 2007–2012 NHANES cycles demonstrate that cotinine levels (a biomarker for nicotine exposure) do not differ by brand, tar group, or menthol status, suggesting that design features may have a limited capacity in reducing toxicant exposure³. In compliance with Framework Convention on Tobacco Control, countries have removed many outward indicators of distinctions among brands through standardized packaging and large pictorial health warnings⁴. France is no exception, having implemented these in 2017⁵. As an EU Member State, manufacturers in France must also comply with the 10-1-10 standard, which was required under the Tobacco Products Directive (2014/40/EU), which came into force in 2014, capping tar, nicotine, and CO emissions⁶. Filter ventilation is a key means of achieving compliance. Despite the lack of evidence supporting harm reduction, and the subsequent banning of misleading descriptors, manufacturers continue to market products in a way that may perpetuate misconceptions^7-13.

Cigarette packaging, in its display of certain color-schemes, has been found to be associated with specific product design characteristics. In fact, less saturated and brighter packaging has demonstrated consistency with greater filter ventilation¹⁴. In recognition of these seemingly deceptive marketing tactics, understanding how filter ventilation and other design features shape consumer perceptions and experiences is important. A series of studies, primarily but not exclusively in the US, have explored the relationship between specific cigarette design features, including filter ventilation, and consumer behaviors and perceptions^15-24. However, little work has been done with respect to cigarette sensory perceptions in a plain-packaging context and an emissions cap.

Several studies have demonstrated evidence that highly ventilated cigarettes are linked to perceptions of lower health risk^18-22, though not all are in support of this finding²³. Furthermore, cigarette brands believed to be ‘light’ – a descriptor commonly used in correspondence with ventilation – have been perceived by consumers as less harmful than regular cigarettes^16,17. Filter ventilation also contributes to sensory experiences, such that cigarettes with greater ventilation are largely perceived as less intense²³, smoother^15,20,23, and lighter^2,15. Additionally, in a US sample of young adults, cork-tipped ventilated cigarettes were situated as the best tasting and most attractive when compared to unfiltered and white-tipped ventilated cigarettes, suggesting that sensory experiences might also vary between types of vented cigarettes²¹. Some studies, however, have published null or dissimilar findings, with ventilation not being associated with perceptions of lightness²⁴ or smoothness¹⁵, and being associated with less taste enjoyment²³.

In recognition of these discrepant findings, further evaluation of the impact of product design features on characteristics and risk beliefs of people who smoke is warranted. In this analysis, we aim to explore the relationship between several design features and perceptions of smoothness and harm, as well as intention to quit smoking. We hypothesized that those using products with higher filter ventilation would perceive them as less harmful and smoother.

METHODS

RESULTS

Table 1 presents the sociodemographic and behavioral profile of the survey respondents for whom brand characteristics data were available. Most of the sample was aged ≤44 years, male, of low to moderate education level, smoked ≤10 cigarettes per day, and perceived themselves as in at least good health. Marlboro was the most commonly reported brand family at nearly a quarter of respondents. The typical cigarette used by respondents had on average 10 mg tar, 0.7 mg nicotine, 28% filter ventilation, was 83 mm long, and contained 815.5 mg tobacco filler.

We explored relationships among the cigarette design features. As anticipated, filter ventilation was strongly negatively related to tar (r= -0.42, p<0.001), while dry weight was positively associated with tar (r=0.29, p<0.001). Interrelationships among other measures are shown in Supplementary file Table S2.

Supplementary file Tables S3A and S3B show mean values for cigarette design features by participant demographics and outcome variables, along with associated ANOVA tests. Pressure drop, ventilation, weight, and length showed consistent relationship to multiple person-level factors. Compared with older age groups, individuals aged 18–34 years reported utilizing cigarettes that on average had lower open pressure drop (82.97 mmWG, p=0.002), closed pressure drop (87.32 mmWG, p<0.001), ventilation (28.0%, p<0.001), overall length (83.07 mm, p<0.001), and wet tobacco weight (832.07 mg, p=0.001). Filter length (23.25 mm, p<0.001) and dry tobacco weight (614.57 mg, p=0.05) were lowest in cigarettes reported by those aged 35–44 years, and diameter (7.68 mm, p=0.02) was lowest in cigarettes reported by those aged ≥55 years. Cigarettes reported by male respondents tended on average to have lower closed pressure drop (87.6 mmWG vs 100.7 mmWG, p<0.001), ventilation (28.5% vs 31.9%, p<0.001), filter length (23.3 mm vs 23.9 mm, p=0.01) and overall length (83.7 mm vs 84.9 mm, p=0.01), as well as larger diameter (7.8 mm vs 7.7 mm, p=0.001). Individuals who report smoking ≤10 cigarettes per day, utilize cigarettes that, on average, have a lower diameter (7.71 mm, p=0.01), wet tobacco weight (833.12 mg, p=0.001), and dry tobacco weight (617.01 mg, p=0.01) when compared with those who smoke more, 11–20 and 21 cigarettes per day. Cigarettes reported by individuals with a low level of education tended to have lower open pressure drop (84.28 mmWG, p=0.03) compared to those reported by individuals with moderate or high level education.

With respect to outcome measures, perceived own-brand harshness was significantly associated with closed pressure drop, ventilation, overall length, and filter length. Perceived harmfulness of own brand was associated with open pressure drop, diameter, wet weight, and dry weight. Perception that smooth/ultra cigarettes are less harmful was associated with wet weight. Intention to quit was not associated with any cigarette design features.

Supplementary file Table S3C displays mean values for tar, nicotine, and CO (ppm), by participant demographics and outcome variables, along with associated ANOVA tests. Interestingly, cigarettes reported by those who perceive themselves as being in poor or fair health tended to emit more CO than cigarettes reported by individuals who perceive themselves as being in at least good health (9.24 vs 9.01, p=0.02).

Table 2 presents results of a stepwise logistic regression model of factors associated with the perception that own brand was smoother than other brands. Initial bivariate analyses indicated that closed pressure drop, ventilation, overall length, and filter length were associated with this outcome. However, stepwise results indicate that only filter length should be retained in the model, and that each 1 mm increase in filter length is associated with 1.11 times greater odds of perceiving own brand as smoother than other brands (p<0.001). Those who reported poor or fair health were significantly less likely to report that their own brand was smoother (AOR=0.66, p=0.02). Age, gender, education level, and cigarettes per day were not significantly related to the perception of own brand as smoother.

Table 3 presents results of a logistic regression model of factors associated with the perception that smooth/ultra cigarettes are less harmful. Initial bivariate analyses indicated that wet tobacco weight was associated with this outcome, but logistic modeling suggests that it was not significantly associated (p=0.16), after controlling for other factors. Younger respondents (18–34 years, AOR=2.69; 35–44 years, AOR=2.37) were significantly more likely than those aged ≥55 years to believe smooth/ultra cigarettes were less harmful. Those who reported poor or fair health (AOR=1.70, p=0.002) and males (AOR=1.88, p<0.001) were also significantly more likely to believe that smooth/ultra cigarettes were less harmful.

Table 4 presents results of a stepwise logistic regression model of factors associated with the perception that own brand was less harmful than others. Initial bivariate analyses indicated that open pressure drop, diameter, wet weight, and dry weight were associated with this outcome, but logistic modeling suggests that only open pressure drop was significantly associated (p=0.02), after controlling for other factors. Specifically, for every 1 mmWG increase in pressure drop, odds of believing own brand was safer increased by 1.02 times. Those aged 45–54 years were significantly less likely than those aged 55 years to believe their brand cigarettes were less harmful. Male respondents were nearly twice as likely to believe their own brand was less harmful (AOR=1.89, p=0.02). Finally, those who reported poor or fair health were also significantly more likely to believe their own brand was less harmful (AOR=1.76, p=0.03). Supplementary file Table S4 reports findings for intention to quit; ventilation was not independently associated with quit intentions.

Table 5 presents the results of a mediation analysis exploring whether perception that own brand was smoother mediates relationships between product design features and perception that own brand was less harmful, while adjusting for demographic factors. Perceiving one’s own brand as smoother fully mediated this relationship for several product design features, including ventilation, closed pressure drop, product length, and filter length. Mediation was strongest with filter length as the predictor, as demonstrated by the indirect effect of 0.13. Additionally, perceiving one’s own brand as smoother was found to partially mediate the relationship between open pressure drop and perceiving one’s own brand as less harmful.

DISCUSSION

While previous studies have consistently demonstrated associations between filter ventilation and odds of perceiving one’s own brand as smoother^15,20,23 and less harmful^18-20,22, our results suggest that other design features consistent with the presence of ventilation may be influencing these perceptions. Adjusting for other factors, own brand cigarette filter length and open pressure drop were significantly associated with greater odds of perceiving one’s own brand as smoother and safer than other brands, respectively.

Consistent relationships between perceived health status and product beliefs are not well documented in the literature. We found health status to be significantly associated with three of our primary outcomes. Specifically, those who reported being of poor or fair health were more likely to perceive smooth/ultra and their own cigarettes as less harmful, as well as less likely to perceive their own brand as smoother. We also noted that those in poor or fair health smoked cigarettes with higher CO emissions on average. Still, it is difficult to extrapolate the direction of effect within these cross-sectional data.

Additionally, we found male respondents to be more likely to perceive smooth/ultra and their own cigarettes as less harmful. Though few studies have demonstrated significant relationships between gender and product beliefs, results from a study evaluating risk perceptions of filter ventilation in a sample of US adults suggest that females are more likely to perceive their own brand as less harmful²³. This study also observed the odds of perceiving one’s own brand as smoother to be greater in older individuals²³. While we did not observe a relationship between age and perceptions of own brand smoothness, we did find those aged 45–54 years to have lower odds of perceiving their own cigarettes as less harmful, and those aged 18–44 years to have greater odds of perceiving smooth/ultra cigarettes as less harmful, compared to those aged ≥55 years. Our findings are consistent with a previous study by Elton-Marshall et al.¹⁶, which found individuals aged ≥55 years to be more likely to perceive their own brand as less harmful.

There are several factors that might explain the observed differences between previous and current findings. First, cigarette marketing in France has been limited over the past two decades due to restrictions implemented through the Framework Convention on Tobacco Control and the European Union Tobacco Products Directive^4,6. Additionally, due to requirements for France to comply with the EU 10-1-10 standard, design features are expected to differ from products manufactured in countries where such regulations are not in place⁶.

CONCLUSIONS

Despite the banning of many deceptive descriptors, cigarettes continue to be manufactured in ways that perpetuate misconceptions. This cross-sectional study underscores the importance of monitoring certain cigarette design features that can inculcate false perceptions of smoothness and safety. Additionally, the current findings suggest that misconceptions may be more prominent in certain populations, which can be important in shaping more targeted communication efforts.