Associarions Between Fine Particulate Matter (PM2.5) and Childhood-Onset Systemic Lupus Erythematosus

Background: Fine particulate matter (PM2.5) has been linked to induction of oxidative stress as well as pulmonary and systemic inammation. We hypothesized that ambient PM 2.5 variation would be associated with the occurrence of childhood-onset systemic lupus erythematosus (cSLE). Methods: We collected data from the Taiwan National health insurance research database and linked these data to the Taiwan Air Quality-Monitoring Database. Children <18 years old, identied from January 1, 2000 were followed up until the rst diagnosis of cSLE was made or until December 31, 2012. The daily average PM2.5 was categorized into four quartile-based groups (Q1-Q4). We measured the incidence rate, hazard ratios (HRs), and 95% condence intervals for cSLE stratied by the quartiles of PM2.5 concentration using Cox proportional hazards models adjusted for age, sex, monthly income, and urbanization. Results:


Introduction
A diagnosis of childhood-onset systemic lupus erythematosus (cSLE) is made when individuals aged less than 18 years develop SLE. [1] cSLE accounts for 10%-20% of all SLE cases. Compared with adultonset SLE, cSLE has a worse clinical course with signi cantly more lupus nephritis, hematological disorders, neurologic disorder, polyarthritis, mucocutaneous involvement, and photosensitivity. [2] Although the etiology of SLE remains unknown, it is multifactorial, including genetic, hormonal, immunologic, and environmental factors. Several environmental factors are reported to be associated with SLE, such as silica exposure, current cigarette smoking, exogenous estrogens, ultraviolet light, solvents, pesticides, heavy metals, and air pollution. [3] There is a growing interest in the role of air pollution on in ammatory diseases, especially concerning particulate matters (PM). Sources of PM are mostly from human activities, including tra c and industrial emissions. [4] Fine PM (with a median diameter < 2.5 mm, PM2.5) are more toxic than other inhalable particles because they can reach deeper areas of the respiratory tract and can be absorbed into the bloodstream through alveolar capillaries, resulting in a regional and even systemic, in ammatory process. [5,6] Previous studies reveal that PM2.5 may be associated with acute and chronic lower respiratory diseases, cerebrovascular diseases, ischemic heart diseases, and lung cancer. [4] Several studies have demonstrated that air pollution enhances the risk of autoimmune diseases in children. Although it is not clearly known what factors play a role in the pathogenesis of cSLE, it has been reported that exposure to SO 2 and O 3 lead to an increase in pediatric rheumatic diseases hospitalizations, and exposure to PM10, NO 2 , and CO may increase the risk of disease activity in cSLE. [6,7] Moreover, maternal exposure to tobacco and air pollutants during pregnancy is associated with cSLE. [7] Recently, a study from Brazil demonstrated that short-term exposure to both indoor and outdoor PM2.5 was associated with increases in airway in ammation and systemic in ammation in cSLE patients. [8] However, these studies only assess the exposure to PM2.5 and the disease activity and hospitalization over a short period of time. There are limited studies examining the association between PM2.5 variation and the incidence of cSLE over a long period of time. Therefore, our objective was to evaluate the effects of air pollution on the risk of developing cSLE in Taiwan from 2000-2012.

Data Source
The data used in the current study were sourced from the Children le, a representative subset of data that includes data from half of all children randomly selected from the year 2000 registry of bene ciaries of the Taiwan National Health Insurance Research Database (NHIRD). The NHIRD was established in March 1995 and includes detailed information, such as outpatient visits, hospital admissions, prescriptions, procedure, and diagnosis of disease, based on the International Classi cation of Diseases, Ninth Revision, and Clinical Modi cation (ICD-9-CM), from 99% of the 23 million enrollees in Taiwan (http://www.nhi.gov.tw/english/index.aspx). The data were analyzed anonymously. This study has been approved by the Institute Review Board of China Medical University Hospital (CRREC-103-048) and complies with the principles outlined in the Helsinki Declaration.

Study population, outcome of interest, endpoints, and confounding factors
We identi ed children < 18 years old from January 1, 2000, to December 31, 2012. Children who had missing data and were diagnosed with SLE before the baseline were excluded. SLE was de ned by at least 3 records of ICD-9-CM code 710.0 made in any diagnosis eld during the inpatient or ambulatory claim process, as our outcome of interest. The Taiwan National Health Insurance (NHI) has classi ed SLE as a catastrophic illness, and the diagnosis of SLE must be con rmed by a board-certi ed specialist and be reviewed and approved by the Taiwan NHI. All participants were followed from baseline until the diagnosis of SLE was made, or patients withdrew from the NHI, or until December 31, 2012. In this study, the mean standard deviation (SD) follow-up duration in SLE patients was 11.2 (2.32) years. The confounding factors were age, sex, urbanization level of residence, and monthly income. Urbanization level was de ned based on population density and was strati ed into four levels, from the highest density (Level 1) to the lowest density (Level 4). Monthly income was classi ed into 4 groups; < NT$14,400, NT$14,400-18,300, NT$18,301-21,000, and > NT$21,000.

Exposure measurement
The Taiwan

Statistical analysis
The demographic categories in the present study included age, sex, urbanization level of the residential area, and the daily average of exposure to air pollutants. To test the distributed difference among daily average concentrations of PM2.5 by quartile and urbanization, a χ2 test was used. The Kaplan-Meier method was used to estimate the proportion of study subjects who did not suffer from SLE during the follow-up period, among the different quartiles of PM2.5 level. The incidence density rate of cSLE (per 100,000 person-years) was counted by each quartile of daily average concentrations of PM2.5. A Cox proportional hazard regression was used to estimate the hazard ratios (HRs) and 95% con dence intervals (CIs) for SLE in the Q2-Q4 levels for air pollutant concentration, compared to the lowest one (Q1). A multivariable model was adjusted for age, sex, monthly income, and urbanization. All analyses were performed using SAS 9.3 (SAS Institute Inc, Cary, NC) and the Statistical Package for the Social Science (Version 15.1; SPSS Inc, Chicago, IL). All statistical results were considered statistically signi cant when 2-tailed P values were < 0.05.

Results
A total of 394 children (0.16%) were newly diagnosed with SLE among a cohort of 244607 children from January 1, 2001 to December 31, 2012. The demographic factors of the study subjects are shown in Table 1. The mean age of participants was 6.09 years (SD, 2.99), and the proportion of boys and girls was 51.8% and 48.2%, respectively. In the present study population, more children lived in higher population density areas (65.3%).  *Chi-square test The urbanization level was categorized by the population density of the residential area into 4 levels, with level 1 as the most urbanized and level 4 as the least urbanized.
The daily average air pollutant concentrations were categorized into 4 groups based on quartiles for each air pollutant.
The incidence rate for SLE increased with PM2.5 exposure concentration, from 4.7 (Q1) to 21.9 (Q4) per 100,000 person-years ( Table 3). The Kaplan-Meier plots ( Fig. 1) with PM2.5 concentration strati ed by quartile showed that patients exposed to higher PM2.5 concentrations had a higher accumulative incidence of SLE than did those exposed to lower PM2.5 concentrations during the 12-year observation period. In the multivariable Cox proportional hazard regression, the adjusted HR for SLE increased with the PM2.5 exposure concentrations from 2.74 to 4.23 compared with that for those exposed to the corresponding concentrations in the Q1 level (Table 3).

Discussion
This is the rst large population study to evaluate the exposure of ambient PM2.5 and the occurrence of cSLE over a long period of time. This longitudinal study showed that higher PM2.5 exposure concentrations increased the incidence rate of cSLE in Taiwanese children and suggests that ambient PM2.5 exposures may be a trigger for the development of cSLE.
Seventy years ago, the historic smog disaster, the 1948 Donora smog, killed 20 people and caused respiratory problems for 6,000 out of the 14,000 people living in Donora. [9] Since then, interest has increased regarding the harmful effects of air pollution. In 1963, the Clean Air Act was established and was last amended in 1990; it requires the Environmental Protection Agency (EPA) to set National Ambient Air Quality Standards (NAAQS) for pollutants considered harmful to public health and the environment.
[10] The World Health Organization (WHO) also challenged governments around the world to improve air quality in their cities to protect peoples' health. [11] However, from the report of WHO, there are still approximately 4.2 million deaths resulting from exposure to ambient air pollution and an additional 3.8 million deaths resulting from exposure to household air pollution, every year. Moreover, several model projections indicate that the contribution of outdoor air pollution to premature mortality could double by 2050. [4] Air pollutants can be found anywhere in the air, both outdoors and indoors. Typically, the environment contains a mixture of gaseous and particulate pollutants. [12] Most air pollutants originate from human activities and emissions of ambient air pollution from regional sources may travel long distances across national borders. [13] To protect air quality in the US, the EPA has mandated air quality standards called NAAQS for the following six air pollutants: ozone (O 3 ), lead (Pb), total suspended particulates (TSP) including PM2.5 and PM10, carbon monoxide (CO), sulfur dioxide (SO 2 ), and nitrogen oxides. These six air pollutants are called "criteria pollutants". [14] An increasing number of epidemiological studies have demonstrated that exposure to air pollutants has harmful effects on cardiovascular and respiratory morbidity and mortality, particularly in children. [15][16][17][18] Children are known to have more adverse health effects to air pollution because of their higher minute ventilation, immature immune system, tendency to spend more time outdoors, and the continuing development of their lungs. [17][18][19][20] PM 2.5 causes more of a burden than other air pollutants because these particles are composed of sulfates, metals, and other toxic substances that are adsorbed into their molecules. [19] The physical and chemical composition and size of airborne particulate matter vary widely with time and space. [20,21] The airborne particulate matter originates from sources such as transportation-related emissions, road/soil dust, biomass burning, and agricultural activities which enter the atmosphere by anthropogenic and natural pathways. [22] PM 2.5 are more toxic because they can reach deeper areas of the respiratory tract and can be absorbed into the bloodstream, resulting in local and systemic in ammation. Exposed to excessive PM 2.5 results in numerous diseases such as asthma, chronic bronchitis, cancer, cardiovascular disease, diabetes, and premature death. [4,[23][24][25][26] For every 10 µg per cubic meter in PM 2.5, all-cause mortality increases by 7.3%. [27] The associations between air pollution and immune-in ammatory responses have been noticed. Exposure to air pollution may cause major autoimmune diseases such as systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), multiple sclerosis (MS), and type 1 diabetes mellitus (T1DM).
[28] Exposure to particulate matter (PM 10 ), sulfur dioxide, nitrogen dioxide (NO 2 ), ozone, and carbon monoxide was found to be associated with high disease activity in juvenile-onset SLE.
[6] A recent Taiwanese study discovered a positive association of NO 2 exposure with the development of SLE in adults. [29] A recent Brazilian study revealed that exposure to inhalable ne particles increases airway in ammation and systemic in ammation in cSLE patients. Although the exact mode of onset and disease progression of SLE remains elusive, the urban-rural difference in prevalence, clustering of disease prevalence around polluted regions, and low concordance rates among monozygotic twins with SLE (around 24%) indicate that environment has a strong impact on SLE. [30] Experimental data strongly suggest that a complex interaction between the exposome (or environmental in uences) and genome (genetic material) produce epigenetic changes (epigenome) that can alter the expression of genetic material and lead to the development of SLE in susceptible individuals. [30] Our study has some limitations. First, since air pollution is a dynamic mixture of different toxicants from natural and anthropogenic sources, including PM, O3, CO, SO2, nitrogen oxides (NOx), and so on, [17] monitoring the concentration of PM2.5 exposure does not fully eliminate the co-effects of mixed air pollutants. Second, since the monitoring stations are xed outdoors, they may not re ect the true exposure level to air pollutants in patients. Third, since this is a retrospective study, we cannot control important confounders such as genetic factors, family history of autoimmune disease, eating habits, leisure activity, sun protection habits, attitudes, body surface area, and cigarette smoking.

Conclusions
In conclusion, exposure to PM2.5 is a risk factor for developing cSLE. Although further studies are required to con rm these associations, our study suggests that awareness, education, and appropriate public policy for better air quality will result in a lower incidence of cSLE and will improve public health. The data were analyzed anonymously and informed consent is not applicable. This study has been approved by the Institute Review Board of China Medical University Hospital (CRREC-103-048) and complies with the principles outlined in the Helsinki Declaration.

Consent for publication:
This manuscript is an original article that has not been previously published and will not be submitted to any other journal. All the authors have read this manuscript and agree that the work is ready for submission, and accept responsibility for the manuscript's contents.
Availability of data and materials: Data available on request due to privacy/ethical restrictions.

Competing interests:
None Figure 1 Kaplan-Meier plot of incidence of Childhood-Onset Systemic Lupus Erythematosus (cumulative incidence rates) in patients with PM2.5 concentration strati ed by quartile.