Combustible cigarettes, heated tobacco products, combined product use, and periodontal disease: A cross-sectional JASTIS study.

BACKGROUND: Combustible cigarettes have detrimental effects on periodontal disease. However, little evidence is available regarding new heated tobacco product (HTP) use and combined product use (both combustible cigarettes and HTPs). This study aimed to examine the association of combustible cigarettes, HTPs, and combined product use with periodontal disease simultaneously.
MATERIALS AND METHODS: This cross-sectional study was conducted using data from the 2019 arm of the longitudinal Japan "Society and New Tobacco" Internet Survey. Combustible cigarette users, HTP users, combined product users, never-users, and former users' data were separately obtained. In the present study, the primary outcome was self-reported periodontal disease. We estimated adjusted prevalence ratios (PRs) and confidence intervals (CIs) using multivariable modified Poisson regression analysis after adjusting for 12 confounders.
RESULTS: Of the 10,439 JASTIS respondents, the numbers of users of combustible cigarettes only, HTPs only, and both products were 1,304, 437, and 1,049, respectively. Compared with never-users, HTP use was significantly associated with the prevalence of self-reported periodontal diseases (PR 1.43, 95% CI 1.03-1.62). Moreover, former users, combustible cigarette users, and combined product users also showed significant associations (PR 1.56, 95% CI 1.35-1.80; PR 1.29, 95% CI 1.03-1.62; and PR 1.55, 95% CI 1.20-1.99, respectively).
CONCLUSIONS: Users of HTPs, combustible cigarettes, and combined products as well as former users were all significantly associated with a higher prevalence of periodontal diseases compared to never-users.

Introduction

Smoking is a global public health problem, as it is a leading risk factor for cancer and chronic respiratory diseases, and was the second leading risk factor for early death and disability worldwide in 2015 [1]. Because of the heavy disease burden due to smoking, tobacco control has been an important issue among healthcare professionals and policymakers [1, 2].

Recently, heated tobacco products (HTPs, also known as “heat-not-burn” tobacco devices) have appeared as a new way to “smoke” tobacco [3]. The first of the new products, IQOS, was launched in Italy and Japan, and the markets have been growing globally [4, 5]. Tobacco industries advertise that HTPs are less harmful to health than combustible cigarettes, presenting HTPs as appealing alternatives to use for smoking cessation [3, 5, 6]. However, recent studies indicate that HTPs are not always safer than combustible cigarettes. It has been reported that HTPs contain more harmful substances than combustible cigarettes [3], and their use has cytotoxic effects on human bronchial epithelium [7]. As HTPs were launched within six years of their invention, nobody knows all their potentially harmful and long-term effects. Therefore, global investigations are needed to clarify the health impacts of HTPs.

Periodontal disease, which is characterized by the loss of periodontal ligament tissue and the collapse of surrounding alveolar bone tissue, is prevalent (7.4%) globally [8]. Periodontal disease causes systemic inflammation and is associated with many systemic diseases, such as cardiovascular diseases, cancers, rheumatoid arthritis, and chronic kidney diseases [9]. Periodontal disease, especially severe chronic periodontal disease (SCP), is the second leading cause of disability-adjusted life-years (DALYs) associated with oral conditions [8]. The estimated global burden of SCP accounted for 3.5 million DALYs (95% uncertainty intervals [UI]: 1.4 to 7.2) with an increasing trend from 1995 to 2015 [8]. Taken together, it is clear that reducing the incidence and the prevalence of periodontal disease is an important public health issue, as reducing the prevalence of the condition may lead to reductions in both the associated systemic diseases and subsequent economic burdens.

Smoking is known to have detrimental effects on periodontal health because it changes periodontal microorganisms in the mouth of a smoker, causing periodontal tissue breakdown [10]. The global smoking-attributed burden of periodontal disease was also high, which accounted for 251,160 DALYs (95% UI: 190,721 to 324,241) among 186 countries in 2015 [11]. Recent studies show that electronic cigarettes (e-cigarettes) and combustible cigarettes are associated with these harmful effects, leading to periodontal disease [12]. However, little evidence is available concerning the relationship between HTPs and periodontal diseases. Furthermore, it remains unclear whether such relationships change depending on the type of tobacco product used, such as combustible cigarettes or HTPs. Therefore, we aimed to examine the association of smoking—according to the types of tobacco products (i.e., combustible cigarettes or HTPs) and combined product use—with the prevalence of periodontal disease using data from a large population sample.

Methods

Study design and setting

For this cross-sectional study, we used data from the 2019 Japanese “Society and New Tobacco” Internet Survey (JASTIS). The JASTIS longitudinal cohort study used a series of internet surveys, from 2015 to 2019, to collect information about the prevalence of new tobacco product use, including HTPs and e-cigarettes, information about the use of conventional tobacco products, and participants’ demographic and socioeconomic data. The resulting data were made available for use by other researchers and collaborative research projects. The 2019 survey data were collected between the 2nd and 28th of February 2019. The online questionnaire was designed such that respondents had to answer each question before they were able to proceed to the next, ensuring all questions were answered. Participants of the JASTIS were recruited via a survey panel provided from a major internet research agency in Japan (Rakuten Insight). The agency keeps about 2.3 million panelists and their socioeconomic status, such as education levels, household income, and marital status. The survey panel comprised those who were recruited initially via services of the Rakuten agency group. For the 2015–2018 baseline survey, participants were randomly sampled from the total panelists from the Rakuten Insight database. The follow-up surveys in other years were conducted from all respondents who previously participated in JASTIS. In the 2019 survey, 9,262 participants were recruited from the follow-up survey. The response rate was 62.5% (9,262 /14,825). In addition to the follow-up survey, the 2019 survey collected new participants (1,738) aged 15–24 from the panel because of the small population compared with other age groups. The additional survey was closed when the target number of respondents who had answered the questionnaire was met. In total, 11,000 respondents participated in JASTIS 2019. Further detailed information regarding the JASTIS is described in the study profile [13].

Inclusion and exclusion criteria

This study included data from all respondents of the 2019 survey except those whose responses were inconsistent with the information they had provided in the earlier surveys (2015 to 2018). On the other hand, this study excluded those whose responses were straight-lining or contained discrepancies. For example, we excluded surveys as straight-lining responses if the respondents chose the same answer number for all questions in a set of questions. We also excluded respondents as responses with discrepancies if they reported an amount of tobacco product use but had indicated that they had never used, or were only former users of, tobacco products. In addition to these exclusion criteria, we performed an attention check for respondents, using the question: "Please choose the second from the bottom." Using this attention check, we excluded respondents who selected responses except the second answer from the bottom.

Measurement of exposure (tobacco product use)

To evaluate the association between past tobacco product use, combustible tobacco use, HTP use, and combined use compared to non-smokers, this study defined the exposures to discriminate each smoking status clearly. Respondents defined as current tobacco product users were asked, “During the past 30 days, have you used each tobacco product? (paper-wrapped cigarettes, roll-your-own cigarettes, nicotine e-cigarettes, non-nicotine e-cigarettes, e-cigarettes with unknown nicotine content, Ploom tech, Ploom tech plus, Ploom S, IQOS, glo, cigars, pipes, chewing tobacco, snuff, and hookahs)”; response options were “Yes” and “No.” Among the current users, types of products were classified as follows: combustible cigarette use (paper-wrapped and roll-your-own cigarettes), HTP use (Ploom tech, Ploom tech plus, Ploom S, IQOS, and glo). We discriminated between the respondents who used only combustible cigarettes, HTP only users, and users of both products. Among non-current users, we separated former users from those who had never used tobacco products, which we determined from the survey data, as shown in S1 Table. Hence, each of the respondents included in our study was classified as one of the following: “never user,” “former user,” “combustible cigarette user,” “HTP user,” or “combined product user.”

Main outcome measures

We defined self-reported periodontal disease as a primary outcome. The JASTIS survey included the question, “Do you have any of the following chronic diseases (for which you have received a diagnosis or are regularly visiting the hospital): hypertension, diabetes, asthma, bronchitis or pneumonia, periodontal diseases, atopic dermatitis, otitis media, heart diseases, stroke, COPD, cancer, depression, or other mental disorders?” The response options were “not affected,” “currently affected and regularly visiting a hospital (clinic),” and “currently affected and without regular (hospital/clinic) visits.” The latter two responses about periodontal diseases were combined and defined as “currently affected by periodontal disease (overall periodontal disease),” and the latter one response was defined as “currently affected by periodontal disease under treatment (periodontal disease under treatment).” This questionnaire consists of the same questions and choices as those in the Comprehensive Survey of Living Conditions in Japan [14].

Covariates

Based on the previous studies regarding the association between combustible cigarettes/e-cigarettes and periodontal disease, we selected age, sex, body mass index (BMI), socioeconomic status (marital status, educational levels, household income), alcohol use [12, 15, 16], routine dental checkups [12, 15], secondhand exposure of combustible cigarettes from others, secondhand aerosols of heated tobacco products from others [12], smoking pack-years [17], and comorbidities (hypertension, diabetes mellitus, heart disease) [17-19] as confounders. In addition to such known confounders, we further selected additional potential confounders (stroke, other tobacco product use) for constructing multivariable models. In total, we selected 12 confounders as covariates. These variables were categorized as follows: age (18–24 years, 25–34, 35–44, 45–54, 55–64, and 65–74), sex (male and female), body mass index, as kg/m2 (≤ 18.4, 18.5–24.9, 25–29.9, ≥ 30), alcohol use during the past 30 days, income (≤ 9,999 US dollars/year, 10,000–39,999, 40,000–79,999, 80,000–119,999, ≥ 120,000, and “unknown”; calculated as 100 JP yen = 1 US dollar), education level (less than high school, high school graduate, college or associates degree, bachelor’s degree, and master’s or doctoral degree), Routine dental checkup, use of other tobacco products, secondhand smoke exposure from others (exposure to smoke/aerosols from combustible cigarettes and/or e-cigarettes or HTPs by others), comorbidities (hypertension, diabetes mellitus, bronchitis or pneumonia, heart diseases, and stroke), and smoking history calculated by pack-years (≤ 5, 6–10, 11–20, 21–30, 31–40, 41–50, ≥ 51) [20].

Statistical analysis

First, we calculated the number and frequency of each variable for use as baseline characteristics. Second, multivariable modified Poisson regression analysis was performed to estimate the prevalence ratio (PR) and confidence interval (CI) for the prevalence of self-reported periodontal disease as the primary analysis, adjusting for the 12 covariates [21, 22]. Additionally, we performed two sensitivity analyses. In the first sensitivity analysis, we changed the definition of exposure from overall periodontal diseases to those under treatment. Furthermore, to validate the confounding selection in the main analysis, we performed a second sensitivity analysis which describes a directed acyclic graph (DAG), and constructed multivariable models based on confounders from the DAG (i.e. age, sex, educational level, routine dental checkup, secondhand exposure to combustible cigarettes from others, secondhand exposure to heated tobacco products or e-cigarettes from others, and smoking pack-years), and confirmed the consistencies of the results. To consider the multicollinearity of variables in the multivariable models, we calculated variance inflation factors (VIF) in both the main and sensitivity analyses. The data were analyzed using STATA version 15.1 (Stata Corp., College Station, TX, USA).

Ethical considerations

We obtained web-based informed consent from all the respondents whose data from the JASTIS study were used in our research. This study was approved by the Institutional Review Board of Osaka International Cancer Institute (No. 1412175183) and the National Institute of Public Health (NIPH-IBRA#12112). The data were anonymized before statistical analyses.

Results

Of 11,000 respondents, 10,439 subjects were included, as shown in Fig 1. The number of current tobacco product users was 2,790; of the current users, 1,304 were combustible cigarette users, 437 were HTP users, and 1,049 were combined product users, respectively. The overall prevalence of periodontal disease was 12.3% (1,279/10,439); periodontal prevalence for combustible tobacco product users, HTP users, and combined product users was 15.3% (200/1,304), 15.1% (66/437), and 19.4% (204/1,049), respectively. In contrast, the prevalence of periodontal disease for never-users was 8.0% (463/5,796). The baseline characteristics are shown in Table 1.

Fig 1

Flow diagram of the study.

Table 1

Demographics of respondents.

	Never user		Former user		Current user						Total
					N = 2,790
					Only combustible cigarettes		Only heated tobacco products		Combined use
	N = 5,796		N = 1,853		N = 1,304		N = 437		N = 1,049		N = 10,439
	n	%	n	%	n	%	n	%	n	%	n	%
Age (years)
18–24	1,602	27.6	62	3.3	74	5.7	60	13.7	114	10.9	1,912	18.3
25–34	879	15.2	77	4.2	58	4.5	43	9.8	48	4.6	1,105	10.6
35–44	925	16.0	307	16.6	235	18.0	97	22.2	233	22.2	1,797	17.2
45–54	897	15.5	443	23.9	385	29.5	116	26.6	295	28.1	2,136	20.5
55–64	822	14.2	501	27.0	367	28.1	93	21.3	245	23.3	2,028	19.4
65–74	671	11.7	463	25.0	185	14.2	28	6.4	114	10.9	1,461	14.0
Sex
Male	2,142	37.0	1,355	73.1	974	74.7	327	74.8	867	82.7	5,665	54.3
Female	3,654	63.0	498	26.9	330	25.3	110	25.2	182	17.3	4,774	45.7
BMI (kg/m2)
≤ 18.4	920	15.9	110	5.9	127	9.7	36	8.2	103	9.8	1,296	12.4
18.5–24.9	4,116	71.0	1,236	66.7	869	66.6	293	67.1	689	65.7	7,203	69.0
25.0–29.9	616	10.6	425	23.0	255	19.6	89	20.4	215	20.5	1,600	15.3
≥ 30.0	144	2.5	82	4.4	53	4.1	19	4.3	42	4.0	340	3.3
Marital status
Single	2,851	49.2	370	20.0	425	32.6	106	24.2	290	27.6	4,042	38.7
Married	2,629	45.3	1,333	71.9	746	57.2	291	66.6	672	64.1	5,671	54.3
Widowed/divorced	316	5.5	150	8.1	133	10.2	40	9.2	87	8.3	726	7.0
30-day alcohol use
Present	2,598	44.8	1,277	68.9	899	68.9	294	67.3	799	76.2	5,867	56.2
Income (USD /year)
≤ 9,999	286	4.9	45	2.4	38	2.9	13	3.0	23	2.2	405	3.9
10,000–39,999	1,286	22.2	452	24.4	322	24.7	78	17.8	185	17.7	2,323	22.3
40,000–79,999	1,678	29.0	632	34.1	443	34.0	143	32.7	386	36.8	3,282	31.4
80,000–119,999	753	13.0	282	15.2	179	13.7	85	19.5	215	20.5	1,514	14.5
≥ 120,000	321	5.5	138	7.5	89	6.8	44	10.1	120	11.4	712	6.8
Unknown	1,472	25.4	304	16.4	233	17.9	74	16.9	120	11.4	2,203	21.1
Educational level
Less than high school	197	3.4	52	2.8	56	4.3	13	3.0	29	2.8	347	3.3
High school graduate	1,689	29.2	528	28.5	414	31.7	132	30.2	272	25.9	3,035	29.1
College or associate’s degree	1,185	20.4	354	19.1	272	20.9	83	19.0	144	13.7	2,038	19.5
Bachelor’s degree	2,396	41.3	831	44.9	512	39.3	186	42.5	553	52.7	4,478	42.9
Master’s or doctoral degree	329	5.7	88	4.7	50	3.8	23	5.3	51	4.9	541	5.2
Routine dental checkup
Present	2,864	49.4	982	53.0	564	43.3	230	52.6	572	54.5	5,212	49.9
Use of other tobacco products
Present					103	7.9	52	11.9	219	20.9	374	3.6
Secondhand exposure
To combustible cigarettes	3,034	52.3	1123	60.6	999	76.6	293	67.0	850	81.0	6,299	60.3
To heated tobacco products or e-cigarettes	1,573	27.1	628	33.9	503	38.6	274	62.7	668	63.7	3,646	34.9
Smoking pack-years
≤ 5					381	29.2	179	41.0	238	22.7	798	7.6
6–10					146	11.2	55	12.6	108	10.3	309	3.0
11–20					284	21.8	72	16.5	191	18.2	547	5.2
21–30					214	16.4	55	12.6	188	17.9	457	4.4
31–40					126	9.7	43	9.8	133	12.7	302	2.9
41–50					72	5.5	19	4.3	84	8.0	175	1.7
≥ 51					81	6.2	14	3.2	107	10.2	202	1.9
Comorbidities
History of hypertension	588	10.1	495	26.7	231	17.7	89	20.4	221	21.1	1,624	15.6
History of diabetes mellitus	150	2.6	189	10.2	106	8.1	37	8.5	99	9.4	581	5.6
History of bronchitis or pneumonia	91	1.6	34	1.8	21	1.6	16	3.7	36	3.4	198	1.9
History of heart diseases	39	0.7	54	2.9	21	1.6	7	1.6	40	3.8	161	1.5
History of stroke	26	0.4	32	1.7	9	0.7	8	1.8	16	1.5	91	0.9
Outcomes
Prevalence of periodontal disease
Overall	463	8.0	346	18.7	200	15.3	66	15.1	204	19.4	1,279	12.3
Under treatment	192	3.3	172	9.3	92	7.1	32	7.3	88	8.4	576	5.5

The results of multivariable modified Poisson regression analyses both in the main and sensitivity analyses were shown in Table 2. After adjustment for the 12 confounders, the PR of combustible products, HTPs, and combined use presented statistically significant associations (PR [95% CI]: 1.29 [1.03–1.62]; 1.43 [1.08–1.88]; and 1.55 [1.20–1.99]) compared with never-users, respectively. Moreover, former users were also significantly associated with the outcome (1.56 [1.35–1.80]). After changing the outcome from the overall prevalence of periodontal disease to that under treatment, similar significant associations were consistently found among former users, combustible tobacco product users, HTP users, and combined product users (PR [95% CI]: 1.85[1.48–2.30]; 1.77 [1.27–2.46]; 1.82 [1.21–2.74]; and 1.82 [1.26–2.63], respectively). In the second sensitivity analysis adjusting for DAG-based confounder, similar association was observed (S2 Table). In both main and sensitivity analyses, all VIFs were less than 2.5; and no problematic multicollinearity was observed (S3 Table).

Table 2

Results of multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment.

	Overall periodontal disease				Periodontal disease under treatment
	PR	95%CI			PR	95%CI
Smoking status
Never user	1.00	(Reference)			1.00	(Reference)
Former user	1.56	1.35	–	1.80	1.85	1.48	–	2.30
Current user
Combustible cigarette use	1.29	1.03	–	1.62	1.77	1.27	–	2.46
HTP use	1.43	1.08	–	1.88	1.82	1.21	–	2.74
Combined use	1.55	1.20	–	1.99	1.82	1.26	–	2.63
Demographics
Age
18–24	0.63	0.47	–	0.84	0.63	0.39	–	1.03
25–34	0.62	0.46	–	0.84	0.65	0.41	–	1.05
35–44	1.00	(Reference)			1.00	(Reference)
45–54	1.51	1.25	–	1.82	1.30	0.96	–	1.76
55–64	2.12	1.77	–	2.55	2.16	1.64	–	2.84
65–74	1.81	1.47	–	2.22	1.73	1.28	–	2.33
Sex
Male	1.00	(Reference)			1.00	(Reference)
Female	1.17	1.03	–	1.33	1.21	0.99	–	1.47
BMI
≤ 18.4	1.02	0.85	–	1.22	0.75	0.54	–	1.03
18.5–24.9		(Reference)				(Reference)
25.0–29.9	0.98	0.86	–	1.12	1.10	0.90	–	1.34
≥ 30.0	1.21	0.96	–	1.52	1.61	1.16	–	2.25
Marital status
Single	(Reference)				(Reference)
Married	1.14	0.98		1.32	1.40	1.10		1.79
Widowed/divorced	1.19	0.97		1.45	1.50	1.09		2.08
30-day alcohol use
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.07	0.95	–	1.19	0.94	0.79	–	1.11
Income (USD /year)
≤ 9,999	1.26	0.95	–	1.66	1.59	1.06	–	2.38
10,000–39,999	1.12	0.98	–	1.28	1.15	0.94	–	1.40
40,000–79,999	1.00	(Reference)			1.00	(Reference)
80,000–119,999	0.81	0.68	–	0.95	0.67	0.52	–	0.88
≥ 120,000	0.78	0.63	–	0.97	0.81	0.59	–	1.09
Secret	0.80	0.68	–	0.94	0.71	0.55	–	0.92
Educational level
Less than high school	1.33	1.02	–	1.73	0.93	0.56	–	1.55
High school graduate		(Reference)				(Reference)
College or associate’s degree	0.80	0.69	–	0.93	0.87	0.70	–	1.09
Bachelor’s degree	0.86	0.76	–	0.97	0.84	0.69	–	1.02
Master’s or doctoral degree	0.80	0.61	–	1.05	0.83	0.55	–	1.25
Routine dental checkup
Absent		(Reference)				(Reference)
Present	1.51	1.35	–	1.68	6.22	4.85	–	7.98
Use of other tobacco products
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.29	1.05	–	1.60	1.13	0.81	–	1.59
Secondhand exposure to combustible cigarettes from others
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.24	1.10	–	1.40	1.08	0.90	–	1.29
Secondhand exposure to heated tobacco products or e-cigarettes from others
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.07	0.95	–	1.20	1.04	0.87	–	1.25
Comorbidities
History of hypertension
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.25	1.11	–	1.41	1.19	1.00	–	1.43
History of diabetes mellitus
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.37	1.18	–	1.60	1.59	1.28	–	1.99
History of bronchitis or pneumonia
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.85	1.45	–	2.35	2.08	1.47	–	2.95
History of heart diseases
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.27	0.98	–	1.63	1.15	0.78	–	1.70
History of stroke
Absent	1.00	(Reference)			1.00	(Reference)
Present	1.16	0.83	–	1.63	1.22	0.72	–	2.08
Smoking pack-years
≤ 5	1.00	(Reference)			1.00	(Reference)
6–10	1.10	0.80	–	1.52	1.26	0.81	–	1.98
11–20	1.05	0.80	–	1.37	0.79	0.51	–	1.22
21–30	0.97	0.74	–	1.28	0.93	0.61	–	1.43
31–40	1.10	0.83	–	1.46	0.96	0.63	–	1.47
41–50	1.31	0.97	–	1.77	1.61	1.07	–	2.42
≥ 51	1.14	0.83	–	1.56	1.09	0.66	–	1.77

Note. PR, prevalence ratio; CI, confidence interval; HTP, heated tobacco products; e-cig, electronic cigarette; BMI, body mass index.

Discussion

Our results suggested that HTP use was significantly associated with the prevalence of self-reported periodontal disease after adjustment of 12 confounders, compared with never-users. Furthermore, former users, combustible tobacco product users, and combined users showed the same association as HTP users. The results were consistent even when the outcome was “active” in the first sensitivity analysis and the DAG-based covariate models were constructed in the second sensitivity analysis.

The growing number of people using HTPs is a critical global public health issue, as is the use of combustible tobacco. The prevalence of HTP use and smoking patterns in Japan were influenced by the early introduction and adoption of HTPs in this country [4]. In fact, according to the 2018 Japanese National Health and Nutrition Survey, male and female HTP users accounted for 30.6% (HTP only 22.1%; combined use of combustible tobacco products and HTPs 8.5%) and 23.6% (HTP only 14.8%; combined use 8.8%) of habitual tobacco product users, respectively [23]. Similar to the Japanese trends, the reported use of HTPs in the United States was up to 2.4% in 2018, and appeared to be increasing [24]. The fact that a growing number of people are using HTPs is a relevant public health problem both in Japan and the United States. On the other hand, concerning periodontal disease—also an important public health issue—evidence of HTPs is scarce, although there have been numerous studies that demonstrate the negative impacts of smoking and the association with periodontal disease [10]. Given these facts, this study clarified the association of HTP use, as well as combustible cigarette use and combined use, with periodontal disease using the “fertile” market of HTPs: the Japanese population [4]. To our best knowledge, this is the first study elucidating this association.

Some potential mechanisms might be considered to explain our results. The first is regarding the cytotoxic effect of HTPs on periodontal tissues. For example, emissions of carbonyl aldehydes from HTPs activate the autoimmune system, which causes a breakdown of periodontal matrix and bone loss [25]. In addition, aerosols containing nicotine accelerate oral myoblast differentiation, resulting in a worsening periodontal environment due to delayed wound healing [25]. Along with those known factors, other potential chemical substances that occur at higher levels in HTPs than combustible tobacco products [3] might contribute to the prevalence of periodontal disease. Our results showed that PRs of HTPs for periodontal disease were comparable to that of combustible cigarettes (1.29 [1.03–1.62] vs. 1.43 [1.08–1.88] for overall periodontal disease; 1.77 [1.27–2.46] vs. 1.82 [1.26–2.63] for periodontal disease under treatment). The second explanation concerns the detrimental effect of the product use on the oral microbiome. Nicotine, as well as smoke from combustible cigarettes, causes the development of dysbiosis in the mouth and periodontal inflammation [26]. Although HTPs contain lower levels of nicotine and some known harmful toxicants than combustible cigarettes [3], habitual use of HTPs might contribute to the destruction of oral microbiomes.

A strength of our study is that we adjusted respondents’ biological features (age, sex, and BMI), socioeconomic status (marital status, alcohol use, income, and educational level), comorbidities (hypertension, diabetes, acute respiratory infections, heart diseases, and stroke), routine dental checkup, and smoking-related conditions (other tobacco product use, secondhand exposure to combustible cigarettes or new tobacco products, and smoking pack-years) in the logistic regression model. Moreover, our study had a large sample size (10,439 included respondents and 1,486 HTP users) and a sufficient number of outcomes (1,279 for overall periodontal disease; 576 for active periodontal diseases) to produce reliable estimations in the multivariable logistic regression models [27]. Despite these strengths, our study had several limitations. First, this survey was based on a self-reported questionnaire. Therefore, there might be an underestimation of exposures [28]. Also, the outcome was defined as visits to the dentist for the treatment of periodontal disease; therefore, localization, severity, or subtypes of the disease could not be discerned. Third, this was an observational study; therefore, residual or unmeasured confounders may exist. Fourth, this study used a cross-sectional design, and could not refer to longitudinal association or causation. Fifth, this survey does not represent the general public in Japan; it is still unclear that our results are fully applicable to the Japanese. Further studies that compensate for the above-mentioned limitations are warranted.

The market for HTPs and the number of products continues to grow. HTPs are now available in more than 30 countries [29], and their use may continue to increase globally. Our results might provide meaningful information for users, healthcare professionals, and policymakers about a potential association of HTPs to oral health.

In conclusion, our study indicated that HTP use, as well as combustible cigarette and combined product use, were associated with a higher prevalence of periodontal diseases compared to non-users after adjusting for 12 confounders. Our results suggest that HTPs might be harmful to oral health, similar to the effects of combustible cigarettes. By exposing the association of HTP product use with periodontal disease, this study may lead epidemiology and public health researchers to examine other implications of the use of these new products. Also, further research is needed to confirm our results and investigate the longitudinal associations between HTP use and periodontal disease.

Definition of exposure.

(DOCX)

Click here for additional data file.

Results of sensitivity analysis: Multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment after adjusting for DAG-based confounders.

(DOCX)

Click here for additional data file.

Variance inflation factors to evaluate multicollinearity of each variable for main and sensitivity analyses.

(DOCX)

Click here for additional data file.

A directed acyclic graph of this study.

(DOCX)

Click here for additional data file.

13 Nov 2020

PONE-D-20-24726

Combustible cigarettes, heated tobacco products, combined product use, and periodontal disease: A cross-sectional JASTIS study

PLOS ONE

Dear Dr. Tabuchi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Despite asking many people, I was only able to secure a single review for your paper.  Fortunately, the one review I have is of high quality, so I have decided to invite major revision based on that single review.

Please submit your revised manuscript by Dec 28 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This manuscript reports associations between heated tobacco product use and self-reported periodontal disease in a large Japanese cross-sectional sample. However, additional details are needed and the analytic methods should be modified for this cross-sectional survey.

1. Kassebaum et al. estimated global prevalence of severe periodontitis at 7.4% not “20-50%” as cited in the paper.

2. The authors should include additional relevant works such as Kassebaum et al. 2017 and Schwendicke et al. 2018 on the global burden of disease in the Introduction and/or Discussion sections.

3. What “(1)… [inconsistencies] with the information they had provided in the earlier surveys (2015 to 2018)” were used to exclude respondents? How many surveys were excluded for reasons (1) or (2)?

4. The authors refer to covariates as “12 confounders” but they have not shown them from the literature or empirically to be confounders. So it is clearer to refer to them as covariates or “potential” confounders.

5. Are there any backdoor correlations induced by overadjustment (ie, adjusting for a covariate that is a result of HTP and a cause of periodontal disease)? A directed acyclic graph (DAG) for hypothesized relationships would be helpful to better understand the hypothesized relationships. For example, is adjusting for exposure to other tobacco product use, secondhand HTP smoke, or diabetes problematic because HTP use relates to THEM which relates to periodontal disease? (in contrast, a confounder would relate TO HTP use and to periodontal disease, not result FROM HTP use.)

6. More details about the design of the cross-sectional study should be provided here rather than just citing Tabuchi et al 2019. Was this a random sample of an existing internet panel? What internet panel provider was used (Rakuten)? What was the response rate?

7. The paper (Tabuchi et al 2019) they cite stated “Respondents of an internet study are not representative of the general population, so we conducted statistical adjustment to account for bias” and “The response rate in the follow-up survey was also problematic, given that non-responders differ in a number of ways from the respondents in the survey.” Were survey weights (eg iterative proportional weights) used to attempt to reduce these biases as in ref#11 (Tabuchi et al 2019)? If so, this manuscript itself needs to provide enough basic info about the study design for readers of this paper to understand the design and the appropriateness of the analyses. If not, why were IPWs not used to try to reduce these biases?

8. Why did the authors use logistic regression models for odds ratios (which are appropriate in case-control studies) instead of the more appropriate log-binomial or Poisson regression models for prevalence ratios which can also be performed with Stata?

9. Add multicollinearity assessment for the models which have some covariates that may be correlated too much with each other.

10. The Results reports “prevalence for combustible tobacco product users, HTP users, and combined product”. This is “periodontal prevalence for combustible…” not prevalence of product use, right?

11. Why would former users have stronger relationship to periodontal disease than current use types? Why does having a routine dental visit relate to periodontal disease – could this be reverse causation? Why do you claim the chance of reverse causation is low in the Discussion?

12. The title of Table 2 should list the response variable (i.e. periodontal disease).

13. An editorial item – a patient can be ambulant or ambulatory; a treatment itself is not ambulant or ambulatory, but for ambulant or ambulatory patients.

References:

Kassebaum et al. JDR 2017

Schwendicke et al. J Clin Periodontol 2018

Tabuchi et al. J Epidemiol 2019

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Stuart Gansky

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

25 Jan 2021

Comments from Reviewer #1:

This manuscript reports associations between heated tobacco product use and self-reported periodontal disease in a large Japanese cross-sectional sample. However, additional details are needed and the analytic methods should be modified for this cross-sectional survey.

Response: We thank Professor Stuart Gansky, for his insightful comments and grateful for the review. We have incorporated changes in the manuscript to reflect of the suggestions provided by the reviewer. The changes were recorded by yellow markers. We hope that the revised manuscript is now suitable for publication.

Comment 1:

Kassebaum et al. estimated global prevalence of severe periodontitis at 7.4% not “20-50%” as cited in the paper.

Response: We appreciate this observation and apologize for the incorrect information. We agree that 7.4% describes the global trends of severe periodontitis. Therefore, we added a new reference (#8), and changed the manuscript below.

Changes:

Introduction

Lines 59–61

Periodontal disease, which is characterized by the loss of periodontal ligament tissue and the collapse of surrounding alveolar bone tissue, is prevalent (7.4%) globally [8].

References

Lines 326–333

8. Kassebaum NJ, Smith AGC, Bernabé E, Fleming TD, Reynolds AE, Vos T, Murray CJL, Marcenes W; GBD 2015 Oral Health Collaborators. Global, Regional, and National Prevalence, Incidence, and Disability-Adjusted Life Years for Oral Conditions for 195 Countries, 1990-2015: A Systematic Analysis for the Global Burden of Diseases, Injuries, and Risk Factors. J Dent Res. 2017;96(4):380-387.

9. Nazir MA. Prevalence of periodontal disease, its association with systemic diseases and prevention. Int J Health Sci (Qassim). 2017;11(2): 72-80.

Comment 2: The authors should include additional relevant works such as Kassebaum et al. 2017 and Schwendicke et al. 2018 on the global burden of disease in the Introduction and/or Discussion sections.

Response: We appreciate the reviewer suggestions of these important studies. As the reviewer instructed, we documented the global burden of disease referring to the suggested two papers.

Changes:

Introduction

Lines 63–67

Periodontal disease, especially severe chronic periodontal disease (SCP), is the second leading cause of disability-adjusted life-years (DALYs) associated with oral conditions [8]. The estimated global burden of SCP accounted for 3.5 million disability-adjusted life-years (DALYs) (95% uncertainty intervals [UI]: 1.4 to 7.2) with an increasing trend from 1995 to 2015 [8].

Lines 73–75

The global smoking-attributed burden of periodontal disease was also high, which accounted for 251,160 DALYs (95% UI: 190,721 to 324,241) among 186 countries in 2015 [11].

References

Lines 328–333

8. Kassebaum NJ, Smith AGC, Bernabé E, Fleming TD, Reynolds AE, Vos T, Murray CJL, Marcenes W; GBD 2015 Oral Health Collaborators. Global, Regional, and National Prevalence, Incidence, and Disability-Adjusted Life Years for Oral Conditions for 195 Countries, 1990-2015: A Systematic Analysis for the Global Burden of Diseases, Injuries, and Risk Factors. J Dent Res. 2017;96(4):380-387.

Lines 337–339

11. Schwendicke F, Dörfer CE, Meier T. Global smoking-attributable burden of periodontal disease in 186 countries in the year 2015. J Clin Periodontol. 2018 Jan;45(1):2-14.

Comment 3: What “(1)… [inconsistencies] with the information they had provided in the earlier surveys (2015 to 2018)” were used to exclude respondents? How many surveys were excluded for reasons (1) or (2)?

Response:

We apologize to the reviewer for our inaccurate expressions in the manuscript. Those who met reason (1) were respondents who participated in JASTIS two times or more and whose basic information, such as sex, was inconsistent. The number of such respondents was two in 2015–2018. They are a priori excluded; therefore, we should have described the reason (1) as an inclusion criterion. On the other hand, those who met reason (2) were all the excluded participants in the Fig 1, that is the exclusion criteria. The number of excluded respondents due to (2) was 561. To make it clearer, we distinguished the inclusion and exclusion criteria, and changed the manuscript as follows:

Changes:

Methods

Inclusion and exclusion criteria

Lines 110–116

This study included data from all respondents of the 2019 survey, except those whose responses were inconsistent with the information they had provided in the earlier surveys (2015 to 2018). On the other hand, this study excluded those whose responses contained other discrepancies. For example, we excluded surveys if the respondents chose the same answer number for all questions in a set of questions, or those who reported an amount of tobacco product use but had indicated that they had never used, or were only former users of, tobacco products.

Comment 4: The authors refer to covariates as “12 confounders” but they have not shown them from the literature or empirically to be confounders. So it is clearer to refer to them as covariates or “potential” confounders.

Response:

We thank the reviewer for the insightful comment. We totally agree with the reviewer. Following the comment, we clearly distinguished the known confounders and potential confounders, and cited new references, which account for the confounding factors.

Changes:

Methods

Covariates

Lines 148–157

Based on the previous studies regarding the association between combustible cigarettes/e-cigarettes and periodontal disease, we selected age, sex, body mass index (BMI), socioeconomic status (marital status, educational levels, household income), alcohol use [12,15,16], routine dental checkups [12,15], secondhand exposure of combustible cigarettes from others, secondhand aerosols of heated tobacco products from others [12], smoking pack-years [17], and comorbidities (hypertension, diabetes mellitus, heart disease) [17,18,19] as confounders. In addition to such known confounders, we further selected additional potential confounders (stroke, other tobacco product use) for constructing multivariable models. In total, we selected 12 participant characteristics confounders as covariates. These variables were categorized as follows: age (18–24 years, 25–34, 35–44, 45–54, 55–64, and 65–74), sex (male and female), …

Discussion

Lines 223–224

Our results suggested that HTP use was significantly associated with the prevalence of self-reported periodontal disease after adjustment of potential 12 confounders, …

Lines 287–289

In conclusion, our study indicated that HTP use, as well as combustible cigarette and combined product use, were associated with a higher prevalence of periodontal diseases compared to non-users after adjusting for 12 potential confounders.

References

Lines 352–369

15. Mundt T, Schwahn C, Mack F, Plzer I, Samietz S, Kocher T, et al. Risk indicators for missing teeth in working-age pomeranians - An evaluation of high-risk populations. J Public Health Dent. 2007;67(4):243-249. doi:10.1111/j.1752-7325.2007.00041.x

16. Hanioka T, Ojima M, Tanaka K, Aoyama H. Relationship between smoking status and tooth loss: Findings from national databases in Japan. J Epidemiol. 2007;17(4):125-132. doi:10.2188/jea.17.125

17. Dietrich T, Stosch U, Dietrich D, Kaiser W, Bernimoulin J-P, Joshipura K. Prediction of Periodontal Disease From Multiple Self-Reported Items in a German Practice-Based Sample. J Periodontol. 2007;78(7s):1421-1428. doi:10.1902/jop.2007.060212

18. Similä T, Auvinen J, Timonen M, Virtanen JI. Long-term effects of smoking on tooth loss after cessation among middle-aged Finnish adults: The Northern Finland Birth Cohort 1966 Study. BMC Public Health. 2016;16(1):1-8. doi:10.1186/s12889-016-3556-

19. Dietrich T, Walter C, Oluwagbemigun K, Bergmann M, Pischon T, Pischon N, et al. Smoking, smoking cessation, and risk of tooth loss: The EPIC-Potsdam study. J Dent Res. 2015;94(10):1369-1375. doi:10.1177/0022034515598961

Comment 5: Are there any backdoor correlations induced by overadjustment (ie, adjusting for a covariate that is a result of HTP and a cause of periodontal disease)? A directed acyclic graph (DAG) for hypothesized relationships would be helpful to better understand the hypothesized relationships. For example, is adjusting for exposure to other tobacco product use, secondhand HTP smoke, or diabetes problematic because HTP use relates to THEM which relates to periodontal disease? (in contrast, a confounder would relate TO HTP use and to periodontal disease, not result FROM HTP use.)

Response:

We thank the reviewer for the insightful comment. Although our selection of confounders in the multivariable models was based on the previous studies, it is important to avoid biased estimation due to unnecessary variable adjustment. Following the reviewer’s instruction, we described a DAG (S1 Fig).

In our study, secondhand smoking and secondhand aerosols were measured as passive secondhand exposure from others. Therefore, we consider them as confounders, not intermediate factors.

As described in the DAG, diabetes, stroke, and other tobacco use may be intermediate factors. Furthermore, respiratory infection, hypertension, and heart disease may be colliders. To estimate the possible effect of overadjustment bias and collider stratification bias on the main analysis, we performed an additional sensitivity analysis where confounders were selected based on the DAG and checked the consistency of the results in the main analysis. The results were almost the same as the main analysis, as described below.

Given these discussions, we changed the following sentences regarding secondhand smoking, secondhand aerosols, description of DAG, and an additional sensitivity analysis in the manuscript, and added a supplementary Figure and a Table (S1 Fig, S2 Table).

Changes:

Methods

Covariates

Lines 151–153

…secondhand exposure of combustible cigarettes from others, secondhand aerosols of heated tobacco products from others [12],

Lines 163–165

…secondhand smoke exposure from others (exposure to smoke/aerosols from combustible cigarettes and/or e-cigarettes or HTPs by others),

Statistical analysis

Lines 173–183

Additionally, we performed two sensitivity analyses. In the first sensitivity analysis, we changed the definition of exposure from overall periodontal diseases to those under treatment. Furthermore, to validate the confounding selection in the main analysis, we performed a second sensitivity analysis which describes a directed acyclic graph (DAG), and constructed multivariable models based on confounders from the DAG (i.e. age, sex, BMI, educational level, routine dental checkup, secondhand exposure to combustible cigarettes from others, secondhand exposure to heated tobacco products or e-cigarettes from others, and smoking pack-years), and confirmed consistencies of the results. To consider the multicollinearity of variables in the multivariable models, we calculated variance inflation factors (VIF) in both the main and sensitivity analyses.

Discussion

Lines 226–228

The results were consistent even when the outcome was “active” in the first sensitivity analysis and the DAG-based covariate models in the second sensitivity analysis.

Supporting information

S2 Table. Results of sensitivity analysis: multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment after adjusting for DAG-based confounders

Overall periodontal disease 　 Periodontal disease under treatment

PR 95%CI PR 95%CI

Smoking status

Never user 1.00 (Reference) 1.00 (Reference)

Former user 1.62 1.40 － 1.86 1.91 1.54 － 2.38

Current user

Combustible cigarette use 1.41 1.13 － 1.77 1.95 1.40 － 2.71

HTP use 1.64 1.24 － 2.16 2.14 1.42 － 3.23

Combined use 1.84 1.45 － 2.34 2.18 1.53 － 3.11

Demographics

Age

18–24 0.63 0.47 － 0.84 0.64 0.39 － 1.05

25–34 0.62 0.45 － 0.84 0.64 0.40 － 1.04

35–44 1.00 (Reference) 1.00 (Reference)

45–54 1.52 1.26 － 1.84 1.31 0.97 － 1.77

55–64 2.27 1.90 － 2.72 2.30 1.75 － 3.03

65–74 2.01 1.65 － 2.45 1.91 1.42 － 2.56

Sex

Male 1.00 (Reference) 1.00 (Reference)

Female 1.13 1.00 － 1.29 1.16 0.95 － 1.40

BMI

≤ 18.4 0.99 0.83 － 1.19 0.73 0.53 － 1.01

18.5–24.9 (Reference) (Reference)

25.0–29.9 1.05 0.92 － 1.19 1.20 0.99 － 1.45

≥ 30.0 1.43 1.13 － 1.81 1.95 1.40 － 2.72

Marital status

Single (Reference) (Reference)

Married 1.15 0.98 － 1.33 1.42 1.11 － 1.82

Widowed/divorced 1.17 0.96 － 1.44 1.48 1.06 － 2.05

30-day alcohol use

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.06 0.95 － 1.18 0.92 0.78 － 1.08

Income (USD /year)

≤ 9,999 1.32 1.00 － 1.74 1.71 1.14 － 2.57

10,000–39,999 1.12 0.98 － 1.27 1.13 0.93 － 1.38

40,000–79,999 1.00 (Reference) 1.00 (Reference)

80,000–119,999 0.78 0.67 － 0.91 0.67 0.52 － 0.87

≥ 120,000 0.85 0.75 － 0.96 0.83 0.61 － 1.12

Unknown 0.78 0.59 － 1.02 0.69 0.53 － 0.90

Educational level

Less than high school 1.39 1.07 － 1.81 1.04 0.63 － 1.71

High school graduate (Reference) (Reference)

College or associate’s degree 0.78 0.67 － 0.91 0.85 0.68 － 1.06

Bachelor's degree 0.85 0.75 － 0.96 0.83 0.68 － 1.00

Master's or doctoral degree 0.78 0.59 － 1.02 0.79 0.53 － 1.19

Routine dental checkup

Absent (Reference) (Reference)

Present 1.53 1.38 － 1.71 6.34 4.95 － 8.13

Secondhand exposure to combustible cigarettes from others

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.24 1.09 － 1.40 1.08 0.90 － 1.30

Secondhand exposure to combustible cigarettes from others

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.09 0.96 － 1.22 1.05 0.88 － 1.27

Smoking pack-years

≤ 5 1.00 (Reference) 1.00 (Reference)

6–10 1.03 0.75 － 1.41 1.12 0.72 － 1.75

11–20 0.94 0.73 － 1.23 0.68 0.44 － 1.05

21–30 0.89 0.68 － 1.17 0.84 0.55 － 1.27

31–40 1.02 0.77 － 1.36 0.88 0.57 － 1.34

41–50 1.22 0.91 － 1.65 1.46 0.97 － 2.18

≥ 51 1.09 0.80 － 1.49 　 1.01 0.62 － 1.63

Note. DAG, directed acyclic graph; PR, prevalence ratio; CI, confidence interval; HTP, heated tobacco products; e-cigarette, electronic cigarette; BMI, body mass index

Comment 6: More details about the design of the cross-sectional study should be provided here rather than just citing Tabuchi et al 2019. Was this a random sample of an existing internet panel? What internet panel provider was used (Rakuten)? What was the response rate?

Response:

We thank the reviewer for this critical observation. We have provided the detailed information about the sampling method, the survey panels, and the response rate.

Changes:

Methods

Study design and setting

Lines 95–107

Participants of the JASTIS were recruited via a survey panel provided from a major internet research agency in Japan (Rakuten Insight). The agency keeps about 2.3 million panelists and their socioeconomic status, such as education levels, household income, and marital status. The survey panel comprised those who were recruited initially via services of the Rakuten agency group. For the 2015–2018 baseline survey, participants were randomly sampled from the total panelists from the Rakuten Insight database. The follow-up surveys in other years were conducted from all respondents who previously participated in JASTIS. In the 2019 survey, 9,262 participants were recruited from the follow-up survey. The response rate was 62.5% (9,262 /14,825). In addition to the follow-up survey, the 2019 survey collected new participants (1,738) aged 15–24 from the panel because of the small population compared with other age groups. The additional survey was closed when the target number of respondents who had answered the questionnaire was met. In total, 11,000 respondents participated in JASTIS 2019.

Comment 7: The paper (Tabuchi et al 2019) they cite stated “Respondents of an internet study are not representative of the general population, so we conducted statistical adjustment to account for bias” and “The response rate in the follow-up survey was also problematic, given that non-responders differ in a number of ways from the respondents in the survey.” Were survey weights (eg iterative proportional weights) used to attempt to reduce these biases as in ref#11 (Tabuchi et al 2019)? If so, this manuscript itself needs to provide enough basic info about the study design for readers of this paper to understand the design and the appropriateness of the analyses. If not, why were IPWs not used to try to reduce these biases?

Response:

This study was designed to evaluate the cross-sectional association between HTP use and periodontal disease among JASTIS 2019. In this study, we focused on evaluating the association, not description; representativeness is essential in this case. In addition, IPW is calculated using data from the 2016 Comprehensive Survey of Living Conditions, which is different in 2019. For these reasons, we did not use IPW adjustment. It is indeed true that the participants of JASTIS 2019 do not represent the general public in Japan, and the limitation of our study is this lack of generalizability. We added the following sentences in the Discussion section.

Changes:

Discussion

Lines 279–280

Fifth, this survey does not represent the general public in Japan; it is still unclear that our results are fully applicable to the Japanese.

Comment 8: Why did the authors use logistic regression models for odds ratios (which are appropriate in case-control studies) instead of the more appropriate log-binomial or Poisson regression models for prevalence ratios which can also be performed with Stata?

Response:

In this study, we used logistic regression models and presented odds ratios for outcome measures because it is commonly used in the epidemiological and clinical research. However, as the reviewer pointed out, odds ratio for not rare outcomes may induce overestimation of the risk. Therefore, we conducted a modified Poisson regression analysis and estimated prevalence ratios for each covariate. The results were almost the same as the logistic regressions. Given this analysis and the reviewer’s kind instruction, we changed all the statistical analyses from using logistic regression models to modified Poisson regression models.

Changes:

Abstract

Lines 27–29

We estimated adjusted prevalence ratios (PRs) and confidence intervals (CIs) using multivariable modified Poisson regression analysis after adjusting for 12 confounders.

Lines 32–36

Compared with never-users, HTP use was significantly associated with the prevalence of self-reported periodontal diseases (PR 1.43, 95% CI 1.03–1.62). Moreover, former users, combustible cigarette users, and combined product users also showed significant associations (PR 1.56, 95% CI 1.35–1.80; PR 1.29, 95% CI 1.03–1.62; and PR 1.55, 95% CI 1.20–1.99, respectively).

Methods

Statistical analysis

Lines 170–173

Second, multivariable modified Poisson regression analysis was performed to estimate the prevalence ratio (PR) and confidence interval (CI) for the prevalence of self-reported periodontal disease as the primary analysis, adjusting for the 12 covariates [21,22].

Results

Lines 205–210

The results of multivariable modified Poisson regression analyses, both in the main and sensitivity analyses, were shown in Table 2. After adjustment for the 12 confounders, the PR of combustible products, HTPs, and combined use presented statistically significant associations (PR [95% CI]: 1.29 [1.03–1.62]; 1.43 [1.08–1.88]; and 1.55 [1.20–1.99]) compared with never-users, respectively. Moreover, former users were also significantly associated with the outcome (1.56 [1.35–1.80]).

Lines 212–215

…similar significant associations were consistently found among former users, combustible tobacco product users, HTP users, and combined product users (PR [95% CI]: 1.85[1.48–2.30]; 1.77 [1.27–2.46]; 1.82 [1.21–2.74]; and 1.82 [1.26–2.63], respectively).

References

Lines 374–377

21. Holmberg MJ, Andersen LW. Estimating Risk Ratios and Risk Differences Alternatives to Odds Ratios. 2021:2020-2021. doi:10.1093/aje/kwi188

22. Zou G. A Modified Poisson Regression Approach to Prospective Studies with Binary Data. 2004;159(7):702-706. doi:10.1093/aje/kwh090

Table 2. Results of multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment

Overall periodontal disease 　 Periodontal disease under treatment

PR 95%CI 　 PR 95%CI

Smoking status

Never user 1.00 (Reference) 1.00 (Reference)

Former user 1.56 1.35 － 1.80 1.85 1.48 － 2.30

Current user

Combustible cigarette use 1.29 1.03 － 1.62 1.77 1.27 － 2.46

HTP use 1.43 1.08 － 1.88 1.82 1.21 － 2.74

Combined use 1.55 1.20 － 1.99 1.82 1.26 － 2.63

Demographics

Age

18–24 0.63 0.47 － 0.84 0.63 0.39 － 1.03

25–34 0.62 0.46 － 0.84 0.65 0.41 － 1.05

35–44 1.00 (Reference) 1.00 (Reference)

45–54 1.51 1.25 － 1.82 1.30 0.96 － 1.76

55–64 2.12 1.77 － 2.55 2.16 1.64 － 2.84

65–74 1.81 1.47 － 2.22 1.73 1.28 － 2.33

Sex

Male 1.00 (Reference) 1.00 (Reference)

Female 1.17 1.03 － 1.33 1.21 0.99 － 1.47

BMI

≤ 18.4 1.02 0.85 － 1.22 0.75 0.54 － 1.03

18.5–24.9 (Reference) (Reference)

25.0–29.9 0.98 0.86 － 1.12 1.10 0.90 － 1.34

≥ 30.0 1.21 0.96 － 1.52 1.61 1.16 － 2.25

Marital status

Single (Reference) (Reference)

Married 1.14 0.98 － 1.32 1.40 1.10 － 1.79

Widowed/divorced 1.19 0.97 － 1.45 1.50 1.09 － 2.08

30-day alcohol use

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.07 0.95 － 1.19 0.94 0.79 － 1.11

Income (USD /year)

≤ 9,999 1.26 0.95 － 1.66 1.59 1.06 － 2.38

10,000–39,999 1.12 0.98 － 1.28 1.15 0.94 － 1.40

40,000–79,999 1.00 (Reference) 1.00 (Reference)

80,000–119,999 0.81 0.68 － 0.95 0.67 0.52 － 0.88

≥ 120,000 0.78 0.63 － 0.97 0.81 0.59 － 1.09

Secret 0.80 0.68 － 0.94 0.71 0.55 － 0.92

Educational level

Less than high school 1.33 1.02 － 1.73 0.93 0.56 － 1.55

High school graduate (Reference) (Reference)

College or associate’s degree 0.80 0.69 － 0.93 0.87 0.70 － 1.09

Bachelor's degree 0.86 0.76 － 0.97 0.84 0.69 － 1.02

Master's or doctoral degree 0.80 0.61 － 1.05 0.83 0.55 － 1.25

Routine dental checkup

Absent (Reference) (Reference)

Present 1.51 1.35 － 1.68 6.22 4.85 － 7.98

Use of other tobacco products

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.29 1.05 － 1.60 1.13 0.81 － 1.59

Secondhand exposure to combustible cigarettes from others

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.24 1.10 － 1.40 1.08 0.90 － 1.29

Secondhand exposure to heated tobacco products or e-cigarettes from others

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.07 0.95 － 1.20 1.04 0.87 － 1.25

Comorbidities

History of hypertension

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.25 1.11 － 1.41 1.19 1.00 － 1.43

History of diabetes mellitus

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.37 1.18 － 1.60 1.59 1.28 － 1.99

History of bronchitis or pneumonia

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.85 1.45 － 2.35 2.08 1.47 － 2.95

History of heart diseases

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.27 0.98 － 1.63 1.15 0.78 － 1.70

History of stroke

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.16 0.83 － 1.63 1.22 0.72 － 2.08

Smoking pack-years

≤ 5 1.00 (Reference) 1.00 (Reference)

6–10 1.10 0.80 － 1.52 1.26 0.81 － 1.98

11–20 1.05 0.80 － 1.37 0.79 0.51 － 1.22

21–30 0.97 0.74 － 1.28 0.93 0.61 － 1.43

31–40 1.10 0.83 － 1.46 0.96 0.63 － 1.47

41–50 1.31 0.97 － 1.77 1.61 1.07 － 2.42

≥ 51 1.14 0.83 － 1.56 1.09 0.66 － 1.77

Note. PR, prevalence ratio; CI, confidence interval; HTP, heated tobacco products; e-cigarette, electronic cigarette; BMI, body mass index

Comment 9: Add multicollinearity assessment for the models which have some covariates that may be correlated too much with each other.

Response:

Thank you for your feedback on this area. We calculated variance inflation factor (VIF), and constructed a Table in the supporting information (S3 Table). As a result, the mean VIF was 1.42, and no variables exceeded 2.5 in VIF. Similarly, we calculated VIF in the second sensitivity analysis, and no VIFs exceeded 2.5.

Changes:

Methods

Statistical analysis

Lines 181–183

To consider the multicollinearity of variables in the multivariable models, we calculated variance inflation factors (VIF) in both the main and sensitivity analyses.

Supporting information

S3 Table. Variance inflation factors to evaluate multicollinearity of each variable for main and sensitivity analyses

Variable VIF

Main analysis 　 Sensitivity analysis

Smoking status

Former user 1.33 1.32

Current user

Combustible cigarette use 2.26 2.2

HTP use 1.39 1.35

Combined use 2.43 2.23

Age

18-24 2.21 2.21

25-34 1.56 1.56

45-54 1.81 1.8

55-64 1.97 1.9

65-74 1.94 1.82

Sex

Female 1.43 1.42

BMI

≤ 18.4 1.09 1.04

25.0-29.9 1.12 1.09

≥ 30.0 1.07 1.09

Marital status

Married 1.87 1.87

Widowed/divorced 1.34 1.34

30-day alcohol use

Present 1.19 1.19

Income (USD /year)

≤ 9,999 1.15 1.15

10,000-39,999 1.44 1.43

80,000-119,999 1.28 1.28

≥ 120000 1.17 1.17

Unknown 1.43 1.43

Educational level

Less than high school 1.1 1.09

College or associate’s degree 1.42 1.42

Bachelor's degree 1.52 1.52

Master's or doctoral degree 1.19 1.19

Routine dental checkup

Present 1.07 1.07

Use of other tobacco products

Present 1.18

Secondhand exposure to combustible cigarettes from others

Present 1.27 1.27

Secondhand exposure to HTPs or e-cigarettes from others

Present 1.33 1.32

Smoking pack-years

6–10 1.36 1.36

11–20 1.66 1.64

21–30 1.58 1.57

31–40 1.4 1.4

41–50 1.25 1.24

≥ 51 1.3 1.3

Comorbidities

History of hypertension

Present 1.26

History of diabetes mellitus

Present 1.13

History of bronchitis or pneumonia

Present 1.04

History of heart diseases

Present 1.09

History of stroke

Present 1.06

Mean VIF 1.42 　 1.45

Note. VIF, variance inflation factors; HTP, heated tobacco products; BMI, body mass index; e-cigarettes, electronic cigarettes;

Comment 10: The Results reports “prevalence for combustible tobacco product users, HTP users, and combined product”. This is “periodontal prevalence for combustible…” not prevalence of product use, right?

Response:

We thank the reviewer for making this observation. The reviewer’s correction is exactly what we would like to present. Therefore, we inserted the term “periodontal” in the following sentence on the Result section.

Changes:

Results

Lines 195–198

The overall prevalence of periodontal disease was 12.3% (1,279/10,439); periodontal prevalence for combustible tobacco product users, HTP users, and combined product users was 15.3% (200/1,304), 15.1% (66/437), and 19.4% (204/1,049), respectively.

Comment 11: Why would former users have stronger relationship to periodontal disease than current use types? Why does having a routine dental visit relate to periodontal disease – could this be reverse causation? Why do you claim the chance of reverse causation is low in the Discussion?

Response:

First, we would like to discuss why former users have a stronger relationship to periodontal disease than current users. We did not obtain the duration and the amount of tobacco product used among former uses, contrary to the current user. Therefore, we adjusted for smoking pack-years for current users only. For this reason, we considered that the point estimates among former users were higher than current users.

Second, we discussed the relationship between routine dental checkups and periodontal disease. As the reviewer pointed out, those who had periodontal disease may consult dentists for routine dental checkup. On the contrary, people who undergo dental checkups are more likely to be diagnosed with periodontal disease by dentists. We considered that this association is bidirectional.

Given the above discussions, we agreed with the reviewer’s indications and changed our consideration that the chance of reverse causation is low because of the presence of the above-mentioned bidirectional associations. Therefore, we changed the manuscript as follows.

Changes:

Discussion

Lines 277–279

Fourth, this study used a cross-sectional design, and we could not refer to longitudinal association or causation, although the possibility of reverse causation was low.

Comment 12: The title of Table 2 should list the response variable (i.e. periodontal disease).

Response:

We agreed with the reviewer and rephrased the title of Table 2, as shown below.

Changes:

Table 2 Results of multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment.

Comment 13: An editorial item – a patient can be ambulant or ambulatory; a treatment itself is not ambulant or ambulatory, but for ambulant or ambulatory patients.

Response:

We appreciate the reviewer’s insightful comment. To avoid the misuse of the terms, we deleted “ambulant” throughout the manuscript, and expressed “periodontal disease under treatment.”

Changes:

Methods

Main outcome measures

Lines 143–145

…and the latter one response was defined as “currently affected by periodontal disease under treatment (periodontal disease under treatment).”

Statistical analysis

Lines 174–175

…we changed the definition of exposure from overall periodontal diseases to those under treatment.

Table 1.

Under treatment 　 192 3.3 　 172 9.3 　 92 7.1 　 32 7.3 　 88 8.4 　 576 5.5

Results

Lines 210–212

After changing the outcome from the overall prevalence of periodontal disease to that under treatment…

Discussion

Lines 255–257

(1.29 [1.03–1.62] vs. 1.43 [1.08–1.88] for overall periodontal disease; 1.77 [1.27–2.46] vs. 1.82 [1.26–2.63] for periodontal disease under treatment).

In addition to the above comments, we have re-numbered the references following the insertion of new references.

Submitted filename: Response_to_reviewer_20210125_clean.docx

Click here for additional data file.

22 Feb 2021

PONE-D-20-24726R1

Combustible cigarettes, heated tobacco products, combined product use, and periodontal disease: A cross-sectional JASTIS study

PLOS ONE

Dear Dr. Tabuchi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

In addition to the clarifications the reviewer requests, please pay particular attention to the open access data requirement.

Please submit your revised manuscript by Apr 08 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have very carefully and thoroughly addressed almost all of the issues from the review of the initial draft in the current revision. However, 3 items were could be addressed a bit more fully.

1. The authors addressed the item about excluding surveys inconsistent with past surveys by writing “This study included data from all respondents of the 2019 survey, except those whose responses were inconsistent with the information they had provided in the earlier surveys (2015 to 2018). On the other hand, this study excluded those whose responses contained other discrepancies. For example, we excluded surveys if the respondents chose the same answer number for all questions in a set of questions, or those who reported an amount of tobacco product use but had indicated that they had never used, or were only former users of, tobacco products.”

Choosing the “same answer number for all questions” is often called “straight-lining” while reporting an amount of tobacco use after indicating they never used would be a discrepancy (or “illogical” or “inconsistent”). Should Figure 1 be referred to here?

Also the term “artificial or unnatural response” in the Figure 1 is peculiar; for the question to detect “artificial response” do the authors mean “attention checks”?

2. The DAG diagram is very helpful to understand the authors thinking with confounders, colliders, and intermediate measures. However, the rationale for BMI relating TO (not FROM) HTP use and TO (not FROM) periodontal disease should be clarified; for example, many people use tobacco products to control their weight (lowering BMI) and periodontal disease might cause people to reduce their caloric (food) intake or cease/reduce eating healthy fresh fruits and vegetables in favor of soft processed foods because of difficulties chewing, which would have BMI as an intermediate factor rather than a confounder.

3. It is not entirely clear why the personal identifying information cannot be removed and the de-identified data placed in a repository or available from the authors upon request after the requestor meets some criteria including attestation that he or she will not attempt to re-identify de-identified data.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Stuart Gansky, DrPH

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

1 Mar 2021

Comments from Reviewer #1:

The authors have very carefully and thoroughly addressed almost all of the issues from the review of the initial draft in the current revision. However, 3 items were could be addressed a bit more fully.

Response: We thank Professor Stuart Gansky again for his insightful comments. We deeply appreciate his review. We have incorporated all the changes in the manuscript to reflect the suggestions provided. The changes have been indicated with yellow markers. We hope that the revised manuscript is now suitable for publication.

Comment 1:

1. The authors addressed the item about excluding surveys inconsistent with past surveys by writing “This study included data from all respondents of the 2019 survey, except those whose responses were inconsistent with the information they had provided in the earlier surveys (2015 to 2018). On the other hand, this study excluded those whose responses contained other discrepancies. For example, we excluded surveys if the respondents chose the same answer number for all questions in a set of questions, or those who reported an amount of tobacco product use but had indicated that they had never used, or were only former users of, tobacco products.”

> Choosing the “same answer number for all questions” is often called “straight-lining” while reporting an amount of tobacco use after indicating they never used would be a discrepancy (or “illogical” or “inconsistent”). Should Figure 1 be referred to here?

> Also the term “artificial or unnatural response” in the Figure 1 is peculiar; for the question to detect “artificial response” do the authors mean “attention checks”?

Response: We appreciate this comment and apologize for the use of confusing words and the inaccurate description of Figure 1.

As the reviewer instructed, “artificial or unnatural response to alcohol and drug use (n = 134)” and “comorbidities (n = 41)” in Figure 1 is what we would like to use to mean “straight-lining” responses; and “inconsistent response to smoking status (n = 326)” indicates responses with a discrepancy. In addition, artificial response to a specific question, “Please choose the second from the bottom,” indicates just the attention checks.

Following the author’s kind instruction, we changed the manuscript and Figure 1 as follows:

Changes:

Methods

Lines 110–120

This study included data from all respondents of the 2019 survey except those whose responses were inconsistent with the information they had provided in the earlier surveys (2015 to 2018). On the other hand, this study excluded those whose responses were straight-lining or contained discrepancies. For example, we excluded surveys as straight-lining responses if the respondents chose the same answer number for all questions in a set of questions. We also excluded respondents as responses with discrepancies if they reported an amount of tobacco product use but had indicated that they had never used, or were only former users of, tobacco products. In addition to these exclusion criteria, we performed an attention check for respondents using the question "Please choose the second from the bottom." Using this attention check, we excluded respondents who selected responses except the second answer from the bottom.

Fig 1. Flow diagram of the study

Comment 2: The DAG diagram is very helpful to understand the authors thinking with confounders, colliders, and intermediate measures. However, the rationale for BMI relating TO (not FROM) HTP use and TO (not FROM) periodontal disease should be clarified; for example, many people use tobacco products to control their weight (lowering BMI) and periodontal disease might cause people to reduce their caloric (food) intake or cease/reduce eating healthy fresh fruits and vegetables in favor of soft processed foods because of difficulties chewing, which would have BMI as an intermediate factor rather than a confounder.

Response: We appreciate the reviewer’s insightful comment.

We can hypothesize that BMI is related TO tobacco use because some people use tobacco to control their weight (BMI → HTP use), and periodontal disease is related TO BMI because of difficulty chewing (Periodontal disease → BMI). Another explanation may be considered, that is, BMI might be related TO HTP use because people lose their weight after using tobacco products (HTP use → lowering BMI), and periodontal disease may be related TO lowering BMI due to chewing difficulties (Periodontal disease → lowering BMI). Taken together, BMI may be either an intermediate variable or a collider; therefore, BMI may not be included in the DAG-based model. We consider that the latter explanation, that BMI is a collider in our DAG, is suitable.

Given this discussion, we changed our DAG and all the results regarding sensitivity analyses of the DAG-based model. Even in the revised DAG-based model, the consistency of the results remains unchanged.

Changes:

Lines 179–186

Furthermore, to validate the confounding selection in the main analysis, we performed a second sensitivity analysis which describes a directed acyclic graph (DAG), and constructed multivariable models based on confounders from the DAG (i.e., age, sex, BMI, educational level, routine dental checkup, secondhand exposure to combustible cigarettes from others, secondhand exposure to heated tobacco products or e-cigarettes from others, and smoking pack-years), and confirmed the consistencies of the results.

S1 Fig. A directed acyclic graph of this study

S2 Table. Results of sensitivity analysis: multivariable modified Poisson regression analysis to estimate the prevalence ratios for overall periodontal disease and periodontal disease under treatment after adjusting for DAG-based confounders.

Overall periodontal disease 　 Periodontal disease under treatment

PR 95%CI PR 95%CI

Smoking status

Never user 1.00 (Reference) 1.00 (Reference)

Former user 1.63 1.41 － 1.88 1.96 1.57 － 2.43

Current user

Combustible cigarette use 1.41 1.12 － 1.77 1.91 1.38 － 2.66

HTP use 1.64 1.24 － 2.17 2.15 1.42 － 3.26

Combined use 1.84 1.45 － 2.34 2.18 1.53 － 3.09

Demographics

Age

18–24 0.62 0.46 － 0.82 0.59 0.36 － 0.98

25–34 0.61 0.45 － 0.83 0.61 0.38 － 0.99

35–44 1.00 (Reference) 1.00 (Reference)

45–54 1.53 1.27 － 1.85 1.34 0.99 － 1.81

55–64 2.28 1.90 － 2.73 2.34 1.78 － 3.08

65–74 2.00 1.64 － 2.44 1.91 1.42 － 2.56

Sex

Male 1.00 (Reference) 1.00 (Reference)

Female 1.12 0.99 － 1.26 1.07 0.88 － 1.29

Marital status

Single (Reference) (Reference)

Married 1.12 0.97 － 1.31 1.36 1.06 － 1.74

Widowed/divorced 1.16 0.94 － 1.42 1.41 1.01 － 1.95

30-day alcohol use

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.05 0.94 － 1.17 0.89 0.76 － 1.06

Income (USD /year)

≤ 9,999 1.31 1.00 － 1.73 1.68 1.12 － 2.54

10,000–39,999 1.12 0.98 － 1.27 1.13 0.92 － 1.38

40,000–79,999 1.00 (Reference) 1.00 (Reference)

80,000–119,999 0.81 0.69 － 0.96 0.66 0.51 － 0.86

≥ 120,000 0.81 0.65 － 1.00 0.84 0.62 － 1.14

Unknown 0.79 0.67 － 0.93 0.68 0.53 － 0.89

Educational level

Less than high school 1.41 1.08 － 1.83 1.06 0.64 － 1.77

High school graduate 1.00 (Reference) 1.00 (Reference)

College or associates' degree 0.78 0.67 － 0.91 0.84 0.67 － 1.05

Bachelor's degree 0.85 0.75 － 0.96 0.82 0.68 － 0.99

Master's or doctoral degree 0.77 0.59 － 1.01 0.77 0.52 － 1.16

Routine dental checkup

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.53 1.37 － 1.70 6.29 4.91 － 8.06

Secondhand exposure to combustible cigarettes

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.24 1.10 － 1.40 1.08 0.90 － 1.30

Secondhand exposure to heated tobacco products or e-cigarettes

Absent 1.00 (Reference) 1.00 (Reference)

Present 1.08 0.96 － 1.22 1.06 0.88 － 1.27

Smoking pack-years

≤ 5 1.00 (Reference) 1.00 (Reference)

6–10 1.02 0.74 － 1.41 1.12 0.72 － 1.75

11–20 0.94 0.72 － 1.23 0.67 0.43 － 1.04

21–30 0.89 0.68 － 1.17 0.84 0.56 － 1.28

31–40 1.03 0.78 － 1.36 0.91 0.59 － 1.38

41–50 1.23 0.91 － 1.66 1.47 0.99 － 2.20

≥ 51 1.10 0.80 － 1.50 1.03 0.64 － 1.67

Note. DAG, directed acyclic graph; PR, prevalence ratio; CI, confidence interval; HTP, heated tobacco products; e-cigarette, electronic cigarette; BMI, body mass index

S3 Table. Variance inflation factors to evaluate multicollinearity of each variable

Variable VIF

Main analysis 　 Sensitivity analysis

Smoking status

Former user 1.33 1.32

Current user

Combustible cigarette use 2.26 2.2

HTP use 1.39 1.35

Combined use 2.43 2.23

Age

18-24 2.21 2.2

25-34 1.56 1.56

45-54 1.81 1.8

55-64 1.97 1.89

65-74 1.94 1.81

Sex

Female 1.43 1.36

BMI

≤ 18.4 1.09

25.0-29.9 1.12

≥ 30.0 1.07

Marital status

Married 1.87 1.86

Widowed/divorced 1.34 1.34

30-day alcohol use

Present 1.19 1.18

Income (USD /year)

≤ 9,999 1.15 1.15

10,000-39,999 1.44 1.43

80,000-119,999 1.28 1.28

≥ 120,000 1.17 1.17

Unknown 1.43 1.43

Educational level

Less than high school 1.1 1.09

College or associate’s degree 1.42 1.42

Bachelor's degree 1.52 1.52

Master's or doctoral degree 1.19 1.19

Routine dental checkup

Present 1.07 1.07

Use of other tobacco products

Present 1.18

Secondhand exposure to combustible cigarettes from others

Present 1.27 1.27

Secondhand exposure to HTPs or e-cigarettes from others

Present 1.33 1.32

Smoking pack-years

6–10 1.36 1.36

11–20 1.66 1.64

21–30 1.58 1.57

31–40 1.4 1.4

41–50 1.25 1.24

≥ 51 1.3 1.3

Comorbidities

History of hypertension

Present 1.26

History of diabetes mellitus

Present 1.13

History of bronchitis or pneumonia

Present 1.04

History of heart diseases

Present 1.09

History of stroke

Present 1.06

Mean VIF 1.42 　 1.48

Note. VIF, variance inflation factors; HTP, heated tobacco products; BMI, body mass index; e-cigarettes, electronic cigarettes;

Comment 3: It is not entirely clear why the personal identifying information cannot be removed and the de-identified data placed in a repository or available from the authors upon request after the requestor meets some criteria including attestation that he or she will not attempt to re-identify de-identified data.

Response: We apologize for our unclear explanation about data availability. In addition to the restriction of ethical guidelines in Japan, our data are available only via our collaborative study framework, as explained in our study profile paper [ref]. Also, we obtained web-based informed consent that the individual data will be used only in the context of our study project and not be used in a public repository. Taken together, we cannot place our data in a public repository.

Ref (cited in Reference 13 in the manuscript)

Tabuchi T, Shinozaki T, Kunugita N, Nakamura M, Tsuji I. Study profile: the Japan “Society and New Tobacco” Internet Survey (JASTIS): a longitudinal internet cohort study of heat-not-burn tobacco products, electronic cigarettes, and conventional tobacco products in Japan. J Epidemiol. 2019;29(11): 444-450. doi: 10.2188/jea.je20180116

Submitted filename: Response_to_reviewer_20210301_clean.docx

Click here for additional data file.

10 Mar 2021

Combustible cigarettes, heated tobacco products, combined product use, and periodontal disease: A cross-sectional JASTIS study

PONE-D-20-24726R2

Dear Dr. Tabuchi,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Stanton A. Glantz

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have thoroughly, carefully, and completely addressed the previous reviews having made revisions to the manuscript.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Stuart A. Gansky

22 Mar 2021

PONE-D-20-24726R2

Combustible cigarettes, heated tobacco products, combined product use, and periodontal disease: A cross-sectional JASTIS study

Dear Dr. Tabuchi:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Stanton A. Glantz

Academic Editor

PLOS ONE