Prolonged Exposure to Air Pollution Increases Periodontal Disease Risk: A Nationwide, Population-Based, Cohort Study

: (1) Background: No association between air pollution and periodontitis has yet been shown. Thus, we merged two nationwide databases to evaluate the risk of periodontitis in Taiwanese residents with long-term exposure to air pollution. (2) Methods: We conducted a nationwide retrospective cohort study using the Longitudinal Generation Tracking Database and the Taiwan Air Quality-Monitoring Database. The daily average air pollutant concentrations were categorized into quartiles (Q1, Q2, Q3, and Q4). We carried out Cox proportional hazards models to compute the hazard ratios of periodontitis, with 95% conﬁdence intervals, in Q2–Q4 of the daily average air pollutant concentrations, compared with Q1. (3) Results: the adjusted HR (95 CI%) for periodontitis in Q2–Q4 increased with increased exposure to SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10 from 1.72 (1.70, 1.76) to 4.86 (4.78–4.94); from 1.89 (1.85–1.93) to 2.64 (2.59–2.70); from 1.04 (1.02–1.06) to 1.52 (1.49–1.55); from 1.61 (1.58–1.64) to 2.51 (2.47–2.56); from 1.48 (1.45–1.51) to 2.11 (2.07–2.15); from 2.02 (1.98–2.06) to 22.9 (22.4–23.4, and from 2.71 (2.66–2.77) to 17.2 (16.8–17.6), respectively, compared to Q1. (4) Conclusions: Residents in Taiwan with long-term exposure to higher levels of air pollutants had a greater risk of periodontitis. data curation, K.-H.C. and C.-L.L.; writing—original draft preparation, H.-J.L.; writing—review and editing, K.-H.C.; visualization, H.-J.L., S.C.-S.T., F.C.-F.L., Y.-C.H., S.-W.C., C.Y.H., R.-H.C., C.-L.L. and K.-H.C.; supervision, K.-H.C.; project administration, K.-H.C., C.Y.H. acquisition, K.-H.C.,


Introduction
In recent decades, air pollution has become an important global issue, being highly correlated with human economic activity and climate change [1]. Many countries have seen rapid industrialization and population density, and there is growing evidence that To explore the risk of periodontitis associated with air pollution in Taiwan, we conducted a nationwide retrospective cohort study using the Longitudinal Generation Tracking Database (LGTD 2000) and the Taiwan Air Quality-Monitoring Database (TAQMD). The LGTD contains the data of two million individuals randomly selected from the National Health Insurance Research Database (NHIRD), which was established in 1995 and contains the comprehensive healthcare data of approximately 99% of Taiwanese residents, with information on outpatient visits, ambulatory care, and inpatient hospitalization, as well as diagnostic codes formatted according to the 9th and 10th revisions of the International Classification of Diseases, Clinical Modification (ICD-9-CM, ICD-10-CM). In 2016, Taiwan's Ministry of Health and Welfare established the Health and Welfare Data Center (HWDC) to standardize the management of available healthcare data. The TAQMD has been maintained by the Taiwan Environmental Protection Agency (EPA) since 1998 and includes the daily concentrations of air pollutants from 74 ambient air quality-monitoring stations across Taiwan. We connected the residential areas of the insured population from the LGTD with the air pollutant information from the nearest TAQMD station. The study was approved by the Central Regional Research Ethics Committee of China Medical University, Taichung, Taiwan (CMUH109-109-REC2-031).

Study Participants
People residing in the vicinity of an air quality-monitoring station and covered by the NHI were selected as the study population. Their healthcare data from January 2005 to 31 December 2016 were then extracted. Each subject was monitored from the index date (2005/01/01) until a new diagnosis of periodontitis was made or until the subject was censored due to loss to follow-up, death, withdrawal from insurance, or the end of followup on 31 December 2017.The index date was defined as the date on which the subjects entered the LGTD. Individuals aged under 18 years were excluded, as were those with missing data on pollutant levels. Those who had ever been diagnosed with periodontitis before the baseline were also excluded.

Exposure Collection and Outcome Measurement
A total of 74 air pollution monitoring stations are located on Taiwan's main island. The location of each monitoring station is determined by the population density. Participants' residential addresses were not available from the TNHI program in compliance with the Personal Data Protection Act. Thus, the residential area of the study subjects was defined by the location of their most common clinic or hospital for cold diagnosis (ICD-9-CM 460). Each resident was assigned a pollutant-exposure level in accordance with the information received from the monitoring station located in the residential area. The air pollutants of interest were SO 2 , CO, NO, NO X , NO 2 , PM 2.5 , and PM 10 . The daily concentrations of air pollution data were measured automatically by the monitoring stations. The outcome event was the diagnosis of periodontitis (ICD-9-CM codes 5233, 5234; ICD-10-CM codes K0520, K0521, K0522, K0530, K0531, and K0532). The endpoint of follow-up was the date of withdrawal from the NHI program, development of periodontitis, or 31 December 2017, whichever occurred first. The daily average air pollutant concentrations were categorized into quartiles, with three cut-off points (25th, 50th, and 75th percentiles) as follows: SO 2 concentration (Q1: <2. 71

Statistical Analysis
In descriptive statistics, the distributions of gender, age, comorbidity, and periodontitis were presented as percentages (%), whereas air pollutants were presented as mean (standard deviation). The chi-squared test was used to test the differences among the quartiles of the daily average concentrations of SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10 at each population density level. The incidence of periodontitis per 1000 person-years was then calculated. We carried out Cox proportional hazards models to compute the hazard ratios (HRs) of periodontitis, with 95% confidence intervals (CIs), in Q2-Q4 of the two-year average air pollutant concentrations, compared with Q1. Multivariable models were adjusted for age, sex, population density level, monthly income, comorbidities of alcohol abuse/dependence, tobacco abuse/dependence, chronic obstructive pulmonary disease, asthma, diabetes, obesity, and rheumatoid arthritis.
Furthermore, we considered each pollutant level as a continuous measure in multivariate analyses and computed the HRs of periodontitis, with 95% CIs in Appendix A.
All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A two-sided p-value < 0.05 was considered significant.

Demographic Characteristics and Air Pollutants of Study Population
A total of 292,263 participants were enrolled in the study. The proportion of females was slightly higher than that of males (55.1% vs. 44.9%), and the mean age at enrollment was 41.1 ± 16.2 years. The proportion of participants who lived in areas with the highest population density was 59.9%. There were 4512 (1.54%) participants who showed alcohol abuse, 47,458 (16.2%) who showed tobacco abuse, and 43,818 (15.0%) with chronic obstructive pulmonary disease. The two-year average concentrations of SO 2 , CO, NO, NO 2 , NOx, PM 2.5 , and PM 10 were 4.42 ± 1.68 ppb, 0.60 ± 0.23 ppm, 8.69 ± 8.77 ppm, 28.6 ± 14.0 ppm, 30.2 ± 7.88 µg/m 3 and 54.3 ± 12.8 µg/m 3 , respectively. At the end of the study, 126,167 participants had experienced periodontitis during the study period. The mean follow-up time was 11.4 ± 2.80 years (Table 1). Median, IQR 13.0 (10.9-13.0) † The population density was categorized into 4 levels, with level 1 as the highest level and level 4 as the lowest level; CO, carbon monoxide; NO, nitric oxide; NO X , nitrogen oxides; SD, standard deviation.

Population Density among the Daily Average Concentration of Air Pollutants
As shown in Table 2, the quartiles of the two-yearaverage concentrations of SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10 showed significant differences among the four levels of population density (p < 0.001). Participants residing in the highest population density were exposed to the highest concentrations of CO (29.2%), NO (30.0%), and NO X (28.9%).

Risk of Periodontitis Based on the Quartiles of Daily Average Air Pollutant Concentrations
The incidence rates and hazard ratios (95% CI) of periodontitis for SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10 are presented in Table 3 and stratified into quartiles. From Q1 to Q4, the incidence rates of periodontitis for SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10 increased from 27.9% to 67.2%, from 24.6% to 55.1%, from 27.8% to53.0%, from 25.1% to 58.6%, from 25.9% to 56.4%, from 20.9% to 62.6% and from 22.1% to 54.5% person-years, respectively. After controlling for age, sex, population density, and comorbidities, the adjusted HR (95 CI%) for periodontitis in Q2-Q4 increased with increased exposure to SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10

Discussion
The present retrospective cohort study was the first to assess the association between long-term air pollutant exposure and periodontitis using a large panel database. The results showed that urban resident exposed to higher level of the investigated air pollutants (Table 2). These findings corroborate those from previous studies showing that air pollutant levels are significantly associated with population density, as higher levels of air pollution are observed in metropolitan areas than in rural areas [41,42]. The possible mechanisms connecting periodontitis and air pollution remain unclear. The air pollutants studied in the present work have been widely investigated and associated with various systemic effects, such as elevated systemic oxidative stress, enhanced blood coagulation, decreased plasma antioxidant capacity, autonomic imbalance, activation of the hypothalamic-pituitary-adrenal axis, and endothelial dysfunction [17,[43][44][45][46][47][48][49][50].
One cross-sectional study on two panels of adult subjects, those with chronic periodontal infections and with healthy periodontal conditions, was conducted to evaluate the connections between particulate matter and high-sensitivity C-reactive protein (hs-CRP); the results suggested that adult patients with chronic periodontitis were more susceptible to PM-associated hs-CRP elevation [51]. Furthermore, the toxicology of air pollution exposure has been found to be closely related to cigarette smoking and environmental tobacco smoke (ETS). Humic-like substances (HULIS), a complex of polyacidic organic compounds formed by organic acids and aerosol particles, have been identified as a chemical component common to cigarette smoke and air pollution particles [52]. Several studies have reported that approximately 3% of ambient air particulate matter (PM), 7-10% of tobacco smoke condensate, and 5% of diesel exhaust is composed of HULIS [52,53]. In animal studies, introduction of HULIS into the lungs from cigarette smoke resulted in phagocytosis and accumulation of intracellular iron, leading to iron sequestration and subsequent deficiency of cellular metals [52]. Consequently, the response to functional metal deficiency includes oxidative stress, activation of cell signaling and transcription factors, release of pro-inflammatory mediators, and apoptosis [54,55]. These reactions further prompted the development of tissue inflammation and fibrosis [54,56].
Moreover, the interaction between ozone and air pollution particles has been proposed to further impact the effects of iron sequestration, disruption of cellular metal homeosta-sis, and tissue inflammation [57,58]. Apart from inflammatory reactions induced by air pollutants and cigarette smoke, pathogenic organisms also take part in the progression of periodontal disease. Porphyromonasgingivalis has been implicated as a major etiological component in the initiation and progression of periodontitis [59]. The ability of the pathogen to colonize and proliferate within the human periodontal pocket depends in part upon the acquisition of iron due to its inability to synthesize the protoporphyrin ring [39]. Characteristic secretions of surface-associated proteases and hemagglutinins have been demonstrated to be capable of degrading structural and defense mechanisms and harvesting heme, iron-containing proteins [39]. The virulence of P. gingivalis can be augmented in the presence of cigarette smoke and hydrogen sulfide, which are commonly seen aerosol particles [60,61]. Laboratory research and animal studies have revealed that purified gingipains and trypsin-like proteinase from P. gingivalis can promote MMP-8 and MMP-3 expression in rat mucosal epithelial cells and upregulate the collagen-degrading ability of some human gingival fibroblasts, causing an imbalance in the proliferative and catalytic activity of human gingival fibroblasts [40].
The present study has several advantages over previous studies. First, we used a nationwide database and recruited a relatively large sample size over an 11-year follow-up period, which yielded strong evidence regarding long-term outcomes in the observed Taiwanese population. Second, as 99% of the Taiwanese population is enrolled in the NHIRD, the bias associated with data collection, region, age, and institution was minimized. The use of the NHIRD also eliminated loss to follow-up and enabled the gathering of data from participants with varied demographic characteristics who were geographically dispersed [70]. Relatedly, the long follow-up period prevented any potential bias. Third, population density was incorporated to demonstrate the geographic connection between the severity of air pollution and periodontitis. However, the present study had some limitations. First, although this study merged two nationwide databases, the exposure levels of indoor air pollutants were not available [71]. Evidence has shown that buildings cannot block outdoor NO 2 molecules efficiently. Thus, it is worthwhile to clarify the impact of indoor and outdoor air pollutants in future studies. Second, two large databases were combined: the NHIRD and TAQMD. The pollution levels in the residential areas of the NHIRD insurants were determined based on the location of their most common clinic or hospital for cold diagnosis (ICD-9-CM 460). Therefore, residents without a cold diagnosis during the study period were not enrolled in this study. Most likely they lived in areas with low levels of pollutants. This may have resulted in an underestimation of PD risk. Third, the NHIRD does not record detailed information on other potential confounders, such as socioeconomic status, family history, systemic comorbidities, smoking or alcohol consumption, and diet factors, which could be considered risk factors for periodontitis. Fourth, we conducted this retrospective nationwide cohort study using random sampling. Although this study had a long follow-up period, underdiagnosis due to PD development after the study period may have resulted in an underestimation of PD risk. Fifth, one of the limitations of LGTD is the lack of SES information. Tobacco abuse/dependence cannot completely reflect light smoking behavior and second-hand smoking. Thus, we considered COPD and asthma as the smoking status of the participants [72][73][74]. However, the covariates included are likely inadequate to control for socioeconomic position, which is a primary confounding factor in this analysis. Sixth, the NHIRD was established in early 1996 and was initially incomplete. Thus, only longitudinal data from January 2005 to December 2016 were selected and analyzed. Bias due to miscoding and misclassification of the ICD may have occurred during the transition phase from paper-based to electronic medical records. Additionally, the ICD-9 information used in this study was restricted to the ICD code only. The definitions of periodontitis and gingivitis have varied significantly between different specialties and timing. Since there were no objective image records, photos of the periodontal disease, or other laboratory data in the NHIRD, we relied on the professional judgment and accurate diagnosis of board-certified dentists. We did not determine the validity of the diagnosis in the present study. Furthermore, there are three sources for potential measurement error: (1) only 74 monitors in the whole study area without research enhanced monitors, and no information about the monitor locations; (2) no detailed residential addresses and moving history available; and (3) only raw measurements without temporal and spatial resolutions from advanced prediction models. Despite all these drawbacks, the data were examined carefully.

Conclusions
The present retrospective cohort study demonstrated an association between air pollution and periodontitis. In particular, residents in Taiwan with long-term exposure to higher levels of air pollutants had a greater risk of periodontitis. However, additional studies are required to investigate whether reduced exposure to air pollutants can lower the risk of periodontitis and to elucidate the underlying mechanism.

Appendix B
The adjusted HR (95 CI%) for periodontitis in Q2-Q4 increased with increased exposure to SO 2 , CO, NO, NO 2 , NO X , PM 2.5 , and PM 10   IR, incidence rate (per 1000 person-years); cHR, crude hazard ratio; aHR, adjusted hazard ratio; CI, confidence interval; The two-year average air pollutant concentrations were categorized into 4 groups based on quartiles for each air pollutant.; Adjusted HR †, adjusted for age, sex, population density, monthly income and comorbidities of alcohol abuse/dependence, tobacco abuse/dependence, chronic obstructive pulmonary disease asthma, diabetes, obesity, and rheumatoid arthritis.