Prevalence, Use Behaviors, and Preferences among Users of Heated Tobacco Products: Findings from the 2018 ITC Japan Survey

Heated tobacco products (HTPs), such as IQOS, glo, and Ploom TECH, with a variety of flavored tobacco-containing inserts, have reportedly achieved a significant market share in Japan. We analyzed data from Wave 1 of the ITC Japan Survey, a nationally representative web survey conducted in February to March 2018 among 4684 adult participants to estimate the prevalence of HTP use, describe characteristics of HTP users, and explore user preferences for HTP device and flavor. The overall prevalence of monthly HTP use was 2.7% (1.7% daily use). Virtually all HTP users were current cigarette smokers (67.8%) or former smokers (25.0%); only 1.0% of HTP users were never smokers. Among HTP users, IQOS was the most frequently reported brand used (64.5%), and menthol was the most common flavor reported (41.5%). IQOS was used more by younger respondents and those who reported daily use, while Ploom TECH was more popular among older respondents and non-daily HTP users. This is one of the first non-industry funded studies to explore the use of HTPs in Japan.


Introduction
Heated tobacco products (HTPs) are devices that heat processed tobacco rather than burn the tobacco directly in order to generate a nicotine aerosol for inhalation. Different HTP devices use different heating sources, including electric energy via a battery or carbon tip that is lit with a match or lighter. The HTP system has three common components: (1) an insert (such as stick, capsule, and Table 1. Characteristics of three major heated tobacco products in Japan.

IQOS Glo Ploom TECH
Device picture

Data Source
Data were from the Wave 1 International Tobacco Control (ITC) Japan Survey, a web-based survey administered by Rakuten Insight (https://insight.rakuten.com/) and collected from February to March 2018. The sampling frame of the survey was an existing Rakuten Insight panel that was nationally representative of Japanese cigarette smokers, HTP users, and non-users. Further quotas based on the region of residence, sex, and age, were applied to ensure the final sample of the survey was proportional to stratum sizes based on Japan census data. Adult residents of Japan (aged 20 (the minimum purchasing age) and older, n = 4684) were sampled as participants of this study. Participants completed an online survey, consisting of questions on HTP and cigarette use, and demographic measures, after eligibility screening. Participants were compensated with a standard Rakuten Insight incentive (points that can be redeemed as cash or merchandise) for a 35-min survey. Both survey procedures and materials were approved by the Office of Research Ethics University of Waterloo.

Sociodemographic Variables
Sociodemographic variables recorded in this study were age, gender, income, and education. Age was categorized into 20-29, 30-39, 40-59, and 60 and older. Participants reported their gender as male or female. Based on Japan's 2016 National Health and Nutrition Survey, annual household income (in Japanese Yen) was defined as follows: low (4,000,000 or less), moderate (4,000,001-6,000,000), high (more than 6,000,000), and refused/do not know [26]. Responses for education were classified as follows: low (junior high school/vocational school/high school), moderate (junior college/technical college), high (undergraduate/postgraduate), and other/refused/do not know.

Patterns of HTP Use and Cigarette Smoking
Participants' HTP use status was assessed by the following question: "How often, if at all, do you currently use heat-not-burn products? These include products such as IQOS, Ploom TECH, and glo". Those who answered "daily", "less than daily, but at least once a week", and "less than weekly, but at least once a month" were considered current HTP users. Details of the survey weighting for "less than weekly, but at least once a month" HTP users are provided in Appendix A. Similarly, to assess cigarette smoking status, the following question was asked: "How often, if at all, do you currently smoke cigarettes, including both factory-made and hand-rolled cigarettes?". Participants who responded "daily", "less than daily, but at least once a week", and "less than weekly, but at least once a month" were considered current cigarette smokers. Those who responded "Not at all; I have quit smoking completely" were considered former smokers; while those who responded "Not at all; I have never been a smoker" were considered never smokers. Exclusive users of HTPs or combustible cigarettes were defined as participants who only use either one of the two products and dual users as participants who use both products at least monthly.
Time-to-first HTP use was examined using the following questions: "How soon after waking do you usually draw your first puff on your heat-not-burn product?" for daily HTP users and "On days that you use heat-not-burn products, how soon after waking do you usually draw your first puff?" for weekly and monthly HTP users. Responses were classified into 5 min or less, 6-30 min, 31-60 min, more than 60 min, and refuse/do not know.
The number of tobacco-containing inserts per day was only assessed for daily and weekly HTP users. For daily HTP users, tobacco-containing inserts per day was assessed by the following question: "On a typical day, how many heat-not-burn tobacco sticks or capsules do you usually use?". For weekly HTP users, we asked the following question: "On a typical week, how many heat-not-burn tobacco sticks or capsules do you usually use?" and divided the response by seven to get the number of tobacco-containing inserts per day.

HTP Device Brand Preferences
Device brand preference was assessed using the following question: "Thinking now about the actual heating device that is used with tobacco sticks or capsules. Which of the following devices have you ever used?". The options following the questions were "IQOS", "glo", "Ploom Tech", and "other (specify)". If participants selected more than one device brand, an additional question was asked: "Which heating device do you use most?" with the same options being given for answer.
Reason for choosing device brand was examined using the following question: "In choosing your heating device, was your choice based on any of the following?", and the participants were instructed to respond for each of the following: (a) It may not be as bad for my health; (b) The price; (c) The taste of the sticks or capsules; (d) The design of the heating device, charging tools, etc. (Please only think about the heating/charging device, not sticks or capsules); (e) The time it takes the device to heat the sticks or capsules; (f) The advertising; (g) It is more available; (h) My friends use this product; and (i) People in the media or other public figures use this product. In each statement, the response options were "yes", "no", "refused", and "do not know" thus participants could select multiple reasons for choosing a device brand.

HTP Flavor Preferences
Flavor preference was evaluated through the following question: "What variety of flavor of tobacco sticks or capsules do you usually use?". In addition to "refused", "other (specify)", "do not know", we initially provided 22 options of tobacco sticks and capsules that were available for purchase in the Japanese market. We re-classified those flavors to tobacco, menthol, mentholated fruity, coffee, and refused/do not know. Further details of the classification of the flavor option for tobacco sticks and capsules are provided in Appendix B.

Statistical Analysis
We presented descriptive statistics of the study population in weighted percentage and 95% confidence interval (CI). Rao-Scott chi-square tests were conducted to both examine the difference between general characteristics of the study population and HTP use by smoking status, along with device brand preference. Normality for the number of tobacco-containing inserts per day was tested with the Shapiro-Wilk test. Due to the non-normal distribution of this variable, a Kruskal-Wallis test was employed to examine differences in the number of tobacco-containing inserts per day between each smoking status or device brand. Multiple comparisons were addressed by using Dunn's test to produce adjusted p-values. Weighted multiple logistic regression analyses were performed to investigate the association between the use of each device brand and demographic characteristics, adjusting for frequency of HTP use. All statistical analyses were performed using svy commands in Stata SE version 14.2 (StataCorp, College Station, TX, USA). All tests were two-tailed and considered significant at p < 0.05. Further details on the weighting and sampling strategy are provided in the ITC Japan Survey Technical Report (https://itcproject.org/files/JP1-1.5_Technical_Report_5June2019.pdf). Table 2 shows the characteristics of the study population stratified by smoking status. The overall prevalence of HTP use was 2.7% (95% CI: 2.4-3.0%), while the prevalence of exclusive HTP use was 0.9% (95% CI: 0.7-1.1%). Among HTP users, 67.8% (95% CI: 63.1-72.2%), 25.0% (95% CI: 20.8-29.8%), and 1.0% (95% CI: 0.4-2.0%) were current smokers, former smokers, and never smokers, respectively. A greater proportion of HTP users were male, aged 40-59, and belonged to the high-income (>6,000,000 Japanese Yen) group. The majority of HTP users who were current smokers and former smokers fell within the high education category (i.e., undergraduate or postgraduate degree).
The number of tobacco-containing inserts per day (median, (interquartile range)) for all HTP users and exclusive HTP users was 10.0 (2.8-15.0) and 10.0 (5.0-18.0). While the number of tobacco-containing inserts per day was not significantly different for HTP users who were former smokers and never smokers (former smokers: 10.0 (5.0-18.0) and never smokers: 15.0 (3.0-20.0)), it was significantly lower for those who also concurrently smoked cigarettes (7.0 (1.4-15.0); both p < 0.05).     There was a statistically significant difference for the number of tobacco-containing inserts per day between Ploom TECH users and IQOS users, along with Ploom TECH and glo users (both p < 0.001).

Prevalence and Pattern of HTP Use
This study adds to the limited body of evidence on HTPs by describing sociodemographic characteristics, product use patterns, device brand and flavor preferences among users of HTPs in Japan, including never smokers, former smokers, and current smokers (dual users). Using a national representative sample of tobacco users and validated survey instruments, we estimated that the overall prevalence of current HTP use among adults in Japan was 2.7% in 2018. To our knowledge, only two other independent studies attempted to estimate the prevalence of current HTP use in Japan. Using repeated online surveys, these studies found there was an increase in the prevalence of current HTP use (defined as use in the past 30 days) from 1.3% in 2015 to 4.7% in 2017 [5,11]. Both estimates, however, were combined percentages for current HTP use and ENDS use. While the prevalence of current ENDS use might be negligible in Japan due to the ban on the sale of ENDS, our study provides the most recent and more precise estimate of the prevalence of current HTP use. Future studies should also investigate the prevalence and pattern of HTP use among adolescents (aged 13 to 19).
Both independent and tobacco industry-funded studies have reported that the prevalence of cigarette smoking is trending downwards in Japan [17,27]. Currently, there are no data on the impact of HTP availability and use on cigarette smoking prevalence. Our data showed that the majority of HTP users are current and former cigarette smokers, with only 1.0% of HTP users who have never smoked cigarettes. Because of the limited number of participants in our survey that are HTP users who have never smoked cigarettes, we were not able to analyze this group in a detailed manner. As a HTP user is still exposed to harmful constituents [2], the uptake of HTP use among never smokers and former smokers is a crucial data point to continue assessing into the future. While the use of ENDS has previously been associated with a greater likelihood of initiating cigarette smoking among youth and young adults [28,29], it is currently unknown whether HTPs will follow a similar trajectory. Conversely, a national study in the Republic of Korea observed similar concurrent cigarette use patterns among HTP users [30]. High rates of dual use among current cigarette smokers, along with lower average tobacco inserts used per day, and longer time-to-first HTP than former or never smokers in this study suggest that HTPs may often serve as complementary products, rather than substitutes for cigarette smoking. Future longitudinal studies may show whether HTP dual use patterns are similar to those observed in other countries for ENDS, and whether dual HTP-cigarette use serve as a transitional behavioral state toward cigarette smoking cessation. Still, substantial dual use of cigarettes and HTPs warrants skepticism toward industry claims of risk-reduction.

Device Brand Preference
Among current HTP users, we found that IQOS was the most commonly used HTP device brand, followed by Ploom TECH and glo, consistent with prior research [11]. In addition to IQOS being introduced earlier than competitor products, PMI's IQOS marketing strategy emphasizes important values within the Japanese culture, such as cleanliness, exclusivity, and a high-tech appearance [20], which has likely played a role in IQOS's popularity. As new generations of HTPs are likely to come about, we plan to examine trends in device brand popularity, using the subsequent waves of data that will be collected for the ITC Japan Survey.
Among current HTP users, the odds of using IQOS were higher among young adults, while the opposite was true for Ploom TECH users, indicating a highly important sociodemographic difference between the two HTP brands. This contrasts with the PMI-funded study that reported IQOS adoption was higher in people aged 44 years and older [31]. While there have been instances where PMI's marketing strategy for IQOS was geared towards young people [32], we did not find evidence that Ploom TECH was marketed to an older age group. Interestingly, no significant associations were observed between gender, annual household income, or education level, and use of a particular HTP device brand. IQOS users had higher odds of using HTPs on a daily basis, while Ploom TECH users had higher odds of using HTPs less frequently (weekly or monthly). Taken together, IQOS and Ploom TECH users appeared to have distinct characteristics.
While the most common reason for choosing glo and Ploom TECH was "perceived reduction in health risk compared to smoking", it was "used by friend" for IQOS. This is a somewhat surprising finding, given that reduced risk claims are a major talking point for PMI's IQOS marketing [33]. The reduced risk claims are primarily based on a reduction in emission of numerous toxicants from HTPs as compared to combustible cigarettes. However, a review of industry documents has shown that some substances, such as propylene glycol, glycidol, acetol, and 2-propen-1-ol, are elevated compared to the combustible cigarette, due to the higher content of humectants in the tobacco filler of the HTP inserts [34]. With IQOS being the most popular HTP device brand and its odds of use higher among young adults, the peer use aspect as a motive for using IQOS is unsurprising as peer influence has been documented to influence young adults with respect to tobacco use behavior [35]. Additionally, "endorsement in the media" as a reason for choosing device brand was significantly higher for IQOS than for glo and Ploom TECH. This corresponds with observations from a previous study in Japan, which reported an exponential increase in internet searches for IQOS immediately following the airing of a single popular television episode, in which a Japanese comedian discussed his IQOS use [11]. Future studies should aim to examine whether these differences are substantiated as the market continues to evolve, and the degree to which individual product marketing and promotional strategies may contribute to users' reasons for adopting a particular HTP brand.

Flavor Preference
Regarding HTP flavor preferences, menthol was the most prevalent flavor preferred by current Japanese HTP users. Likewise, the popularity of menthol flavors in combustible cigarettes has been increasing in Japan, particularly among younger and female smokers [36]. Preference for menthol flavors among smokers has been reported to be inversely associated with age [37]. It is not surprising that HTP users, who are predominantly younger, also have preferences toward this flavor. The high popularity of menthol flavor among HTP users may have negative consequences for public health. Menthol flavor in cigarettes has been associated with increased initiation [37,38], greater nicotine dependence [39], lower quit rates [40], and higher relapse [38,39]. The tobacco industry has long utilized menthol flavor to attract users [39]. Given the wide variety of menthol flavors (and by extension other mentholated flavors, including mentholated fruity) being offered as tobacco-containing inserts, and what is known about the public health effects of menthol, the extensive use of menthol flavor for HTPs is of concern.

Study Limitations
There are several limitations to this study. First, data in this study were self-reported and we were not able to validate it, particularly when it comes to abstinence from cigarette smoking among exclusive HTP users. Second, this study was based on cross-sectional data; thus, causal inference should not be made. Third, even though ENDS is not commercially available in Japan, we did not evaluate the concurrent use of ENDS or ENNDS (electronic non-nicotine delivery system) among HTP users. Lastly, although the sampling and weighting strategies seek to ensure national representativeness, the possibility of excluding participants in a web-based survey cannot be ruled out.

Conclusions
In 2018, the prevalence of current HTP use in Japan was 2.7%, with most HTP users also smoking combustible cigarettes. The most popular brand device among HTP users was IQOS, followed by Ploom TECH and glo. Notable differences in age and frequency of HTP use were observed between IQOS and Ploom TECH users. Menthol was the most common flavor being used by HTP users. Given HTPs' significant presence, tobacco regulation in Japan should also include HTPs and should be tailored to specific groups of HTP users.  (FDN-148477). G.T.F. was also supported by a Senior Investigator Award from the Ontario Institute for Cancer Research.
Conflicts of Interest: K.M.C. has received payment as a consultant to Pfizer, Inc., for service on an external advisory panel to assess ways to improve smoking cessation delivery in health care settings. K.M.C. also has served as paid expert witness in litigation filed against the tobacco industry. G.T.F. has served as an expert witness on behalf of governments in litigation involving the tobacco industry. M.L.G. has received a research grant from Pfizer and served as a member of scientific advisory board to Johnson & Johnson. All other authors have no conflicts of interest to declare.

Appendix A Survey Weighting for "Less than Weekly, but at Least Once a Month" HTP Users
For survey weight, the definition for current HTP users in ITC Japan Survey is daily and weekly HTP users. Meanwhile, the definition for current HTP users in this study is daily, weekly, and monthly HTP users. Monthly HTP users were, therefore, eligible for the survey in exclusive smokers and non-users categories. As a result, they were given the survey weight in accordance with each respective group. The current setting implies that for weight calculation, monthly HTP users "behave" more like HTP non-users rather than like HTP users. Similarly, monthly HTP users who were also cigarette smokers "behave" more closely to exclusive smokers than to dual users.
The additions of monthly HTP users increased variability and thus reduce power for the regression analysis. This, however, is compensated by the increase in sample size coming from including monthly HTP users and inclusion of the frequency of HTP use as one of the covariates for the regression analysis.