Analysis of the Association between environmental-meteorological factors and Acute Suppurative Otitis Media: Further Epidemiological Evidence from Lanzhou, Northwestern China CURRENT STATUS: POSTED

Objective To investigate the possible associations, Methods SPSS software was used to analyze Spearman correlation between ambient environment-meteorological factors and the daily number of patient ASOM visits. The GAM of time series was used to analyze the exposure -response curve, and R software was used to calculate RR.Using above results to discuss the relationship of air pollutions , meteorological factors and morbidity of ASOM. We analyzed the distribution characteristics of air pollutants meteorological factors in 2014 to 2016, and the correlation between air pollutants and meteorological factors. We also examined the lag effects of air pollutants, meteorological factors and the possible differences of air pollutants, meteorological factors on the daily number of ASOM visits that were consulting the hospital for treatment for ASOM over a period of 3 years. The results showed that the daily number of ASOM visits had obvious seasonal differences. The air pollutants positively correlated with the daily number of ASOM visits, and had lag effects on the daily number of ASOM visits, temperature is negatively correlated with the daily number of ASOM visits. RR in the number of ASOM visits increased with air pollution level. There are also some limitations to our study. The data at effect of surface modification of fibroblast growth factor 2R cell membrane on noise-induced inner ear injury and its related mechanism (18YF1WA132); Health Industry Research Program of Gansu Province, Effects of Nitrogen Oxide and Ozone Exposure on Secretory Otitis Media and Related Mechanisms (GSWGKY2018-25); Military Medical Science and Technology Youth Cultivation Program of 2018 (18QNP047).


Introduction
Along with the rapid economic growth and urbanization, environmental pollution has become a serious problem due to the extensive use of fossil fuels. Especially in Lanzhou and other areas where heavy industry development is an important support. The impact of environmental change on health has become a sensitive topic for the public, the media and even the Chinese government and adjacent countries [1].
Globally, scholars used cohort studies, case-control studies, cross-case studies and cross-sectional studies to analyze the correlation between air pollutants and otitis media (OM). There is growing evidence that air pollution exposure is associated with OM, and all studies investigating this associftion found evidence that exposure to at least one pollutant increased the risk of OM [2].
MacIntyre et al. reported a significant positive correlation between NO 2 and OM in the largest birth cohort of 10 European cohorts [3]. Brauer et al. found that there was a correlation between children's exposure to nitrogen dioxide(NO 2 ), PM 2.5 and OM [4]. However, A systematic review by Jones et al. reported in 2012 that living with a smoker increased the risk of people's OM by 62% [5]. Zemek et al. reported lag effects between NO 2 , PM 10 ,O 3 and the hospital OM visits [6].
However, for environmental pollutants, only a few studies have reported the effects of air pollutants on the incidence of AOM in China, and the research object is children. Lu [7] and Zhang [8] studied the correlation between air pollution and children's AOM in Beijing. The results showed that air pollutants increased the incidence of AOM in children, especially SO 2 and NO 2 .
Although there is convincing evidence for an association between environmental tobacco smoke exposure and OM [9], the relationship between ASOM and ambient air pollution exposure is not yet established, and the impact of meteorological factors on ASOM is rarely reported. As one of the most common diseases in children, ASOM mostly secondary to respiratory tract infections, and is the most common cause of antibiotic use in people. It is clinically characterized by earache, ear purulence, tympanic membrane congestion and perforation. If the treatment is not timely, it may lead to tympanic membrane perforation, hearing loss, and even chronic suppurative otitis media, which will bring great economic and social burden to families and health care systems [10,11]. Therefore, it is of great significance to explore the pathogenesis, early prevention and treatment of ASOM. To provide reference for reducing and preventing the occurrence of diseases and rationally arranging medical resources.
In this study, we collected data of ASOM from 2014 to 2016 in Lanzhou, and used time series analysis method to analyze the relationship between environmental-meteorological factors and the daily number of ASOM visits. Hospital of the Joint Logistics Support Unit of the Chinese PLA and Second Affiliated Hospital of Lanzhou University. Both hospitals are Class-Three, and have advanced inspection equipment, experienced clinicians and perfect electronic medical record system. The most important is undertake a large part of the diagnosis and treatment of Otolaryngology diseases in Lanzhou, which ensures sufficient sample size and the reliability of medical record data. They are both comprehensive teaching and researching medical centers.

Materials And Methods
The principal diagnosis of ASOM was based on medical history, physical examination, auxiliary examination. Case data included patient's medical record number, first name, last name, sex, age, date of birth, diagnosis, and excluded ANSOM according to international disease diagnostic criteria. In order to avoid repeated counting, only one visit per individual patient per day was used as daily visit counts. Subsequent follow-ups within 30 days of the initial visit were excluded. All patients have given written consent to participate in the study. All medical interviewers (general practitioners or nurses) were trained to use uniform examination protocols. Medical records and the respective results were confirmed by the supervisors at each hospital. This study was approved by the Gansu Provincial Ethics Committee and the Human Research Ethics Committee of the Ministry of Health.

Descriptive Statistics
Frequency distribution characteristics of meteorological data, air pollution data and disease data were analyzed, including mean, SD, Minimum and Maximum, and 25, 50 and 75 percentiles (P25, Median and P75, respectively). The characteristics of annual and monthly changes of disease data were analyzed. In addition, the disease data were stratified, taking into account the distribution characteristics of different ages (0-14 years old, 15-64 years old and over 64 years old) .

Correlation Analysis
SPSS version 24.0 was used to analyze spearman correlation between ambient environmentmeteorological factors and the daily number of patient ASOM visits.

Generalized Additive Model
Generalized additive model(GAM)has become one of the most widely used methods in studying the effects of meteorological and environmental factors on human health. It is more flexible than other statistical models and has greater advantages in dealing with the complex non-linear relationship between independent variables and strain variables. Due to the daily number of ASOM visits was sparse and typically followed a Poisson distribution [9], the core analysis was a GAM with log link and Poisson error that accounted for smooth fluctuations in daily number of consultations for ASOM patient. A detailed introduction of the GAM has been previously described [10]. GAM was used to analyze the influence of meteorological factors and air pollutants on the daily number of patients with ASOM. Due to the number of hospitalized patients changes nonlinearly with time, spline smoothing function is used to control the long-term trend. At the same time, weekend and holiday effects are taken as dumb variables to control. The degree of freedom of model parameters is judged by Akaike's Information Criterion(AIC). The basic model is: Yt refers to the number of ASOM visits on day t; E (Y) refers to the expected number of ASOM visits on day t; Xt refers to the average daily concentration of pollutants on day t; l is the lag day of air pollutant exposure, which is a commonly used concept in time series analysis. Its meaning is to use the health effect index of that day and the pollutants of the previous n days. Regression analysis was used to study the influence of pollutant concentration on the number of patients visiting ASOM in the past few days. α was residual; β was regression coefficient, which indicated that the relative variable For seasonal analysis, seasonality was differentiated on the basis of heating/non-heating periods. This study divides the seasons according to the meteorological characteristics of Northwest China, it is defined that spring is from February to April (2)(3)(4)(5), the summer is from May to August(5-8), the autumn is from September to November (9)(10)(11), and the winter is from December to next year January(12 − 2).
For the lag effects model, we examined the effect of air pollutants with different lag (L) structures of single-day lag (distributed lag; from L0 to L7) and multi-day lag (moving average lag; L0-1 to L0-14).
In this study ,a lag of 0 day (L0) corresponds to the current-day pollution, and a lag of 1 day refers to the previous-day concentration. In multi-day lag models, L0-14 corresponds to 15 day moving average of pollutant concentration of the current and the previous 14 days [12] .The data of meteorological factors used in the lag model are similar to those of atmospheric pollution.
In this study, R software (version 3.5.0) with mgcv package was used to analyze the exposureresponse relationship between atmospheric pollutants and meteorological factors and the daily number of ASOM visits. The results obtained were expressed as the relative risk (RR) percentage change in the daily number of ASOM visits for per 10 µg/m3(1 mg/m3) increases of air pollutant concentrations. Relative risk (RR) and 95% confidence interval (95% CI) were calculated. The test level was 0.05 (P < 0.05 test had statistical significance, p < 0.01 test had statistical significance).

Regional Survey Of Research
Lanzhou is located in the western part of the Loess Plateau and the eastern part of the Qinghai-Tibet Plateau. Its geographical location is 35 34'-37 07', 102 36'-104 34', high in the West and south, low in the East and north. The Yellow River runs through the whole city from southwest to northeast, forming typical geomorphological features of the valley basins. The urban area is about 35 kilometers long from east to west and 8 kilometers wide from north to south. It is located in the Inland hinterland, in the transition zone between monsoon area and non-monsoon area, and belongs to temperate semi-arid climate.

Quality Control
Medical records are managed and checked in accordance with the national standards of the thirdclass and first-class hospitals. Personnel are responsible for the entry and management of medical records and excluding cases with incomplete information, unclear diagnosis and non-acute suppurative otitis media. Air pollution data and meteorological data are from the national certified automatic atmospheric monitoring system and meteorological observation system. Personnel are responsible for inspection, identification and management. In this study, the data used are aggregated and calculated through data management software, and the analysis, statistics and management of data are strictly controlled.  Table 2 summarizes the distribution of the annual mean atmospheric pressure, temperature, relative humidity and other meteorological factors of Lanzhou during the study period.

Characteristics of Air Pollutants and Meteorological Factors
During our study period, the average atmospheric pressure was 811.3 hPa and the average wind speed was 1.99 m/s. The relative humidity ranged from 21-100%, the mean value was 60.56%, the temperature ranged from − 17℃ to 25 ℃, and the average value was 7.74 C. The average maximum temperature was 18.13 C, the minimum temperature was 5.81 C, and the average daily temperature difference was 12.32 C. respectively, and the minimum number of patients per day for all different groups was 0.
Characteristics of ASOM patients are also summarized in Table 3.

Seasonal Effect
Annual morbidity of ASOM was 916, 875 and 1023 respectively. The results showed that the number of patients in Summer was less than the other three seasons (Fig. 2), The number of patients was the largest in winter, and there was no significant difference between spring and autumn(p = 0.65).

Monthly Effect
The distribution of monthly visits in ASOM showed that the number of visits increased gradually from September to December, but there was an episode in November 2015. It was noted that peak could be seen in December, and a valley in August (168), followed by February (171), and the medium level in March to May, as shown in Fig. 3.      Table 6 shows the multi-day lag effects of environmental variables on the daily number of patient ASOM visits. Italic value represents the strongest multi-day lag effect days.

Associations between Air Pollutants and Meteorological Variables
The results of this study showed that the multi-day lag effects of environmental-meteorological factors were mainly concentrated in 3-6 days. The environmental factors PM2.5, PM 10 , CO and NO 2 we selected were positively correlated with the daily number of ASOM visits, and had the multi-day lag effects on the daily number of ASOM visits. Among them, PM 2.5 and PM 10 had the strongest multiday lag effects on day 13 (L0-13) and day 14 (L0-14) respectively, the strongest multi-day lag effects of NO 2 and CO were on the same day (L0-6), SO 2 and O 3 did not pass the significance test (P > 0.05).
The meteorological factor ATM was positively correlated with the daily number of ASOM visits, while there were negatively correlated with RH, W and T. And ATM, T, W had the multi-day lag effects on the daily number of ASOM visits, the strongest multi-day lag effects of ATM, W, and T were on day 6 (L0-6), day 3 (L0-3) and day 5 (L0-5), respectively. RH passed the significance test only on the 11th day(L0-11). Table 6 Spearman analysis of multi-day lag effects on ASOM by environmental variables  In this study, we found generally linear relationships (monotonic trends) for daily number of ASOM visits associated with CO, NO 2 and T. In addition, we observed basically monotonic increased RR for both CO and NO 2 within these ranges of concentrations (Fig. 4), indicating that CO and NO 2 are significantly associated with the daily number of ASOM visits. we also observed basically monotonic increased RR for NO 2 , but there is a turning point in the process of increasing. RR increased slightly while the concentration of NO2 was less than 50 µg/m3, and increased more while the concentration of NO2 was more than 50 µg/m3. The daily number of ASOM patients was negatively correlated with CO concentration (CO < 1 mg/m3), however, with the increase of CO concentration (> 1 mg/m3), the number of ASOM patients increased. Temperature was negatively correlated with the daily number of ASOM patients.
The curve of PM 10 and O 3 are "N". When the concentration of PM 10 is between 250 µg/m3 and 600 µg/m3, the RR decreases gradually. When the concentration of PM 10 is less than 250 µg/m3 or more than 600 µg/m3, the RR increases gradually.
Moreover, we observed basically monotonic increased RR for both RH and PM 2.5 , within these ranges ( Fig. 4), indicating that PM 2.5 and RH are significantly associated with increased hospital visits of ASOM. The curve of SO 2 was "M" type with two peak threshold ranges. The first RR peak appeared at the concentration of SO 2 of 20 µg/m3, and the second peak appeared at the concentration of NO2 of 55 µg/m3, that means when the SO 2 concentration is higher or lower, the risk of seeing a doctor for ASOM is reduced. The RR was minimum when the NO 2 concentration was 35 µg/m 3 .The curve of PM 10 and RH were "S" and there was a dangerous threshold concentration. ATM was positively correlated with the daily number of ASOM visits. RR increases with the increase of ATM. With the increase of T, the RR value increases first and then decreases. That means when the T is higher or lower, the risk of patients with ASOM will decrease.   Although there are some limitations, it still reflects the acute health effects of air pollution and metrological factors on the incidence of ASOM in exposed population in Lanzhou.

Characteristics Of Environmental-meteorological Factors
The which forms the unique landform and meteorological conditions. Thus, the diffusivity of air pollutants are greatly reduced. In addition, the petrochemical industry, the rapid growth of car ownership and the surrounding fragile ecological environment make air pollution gradually worsen in Lanzhou.

Pathogenic Characteristics Of ASOM
Our study demonstrated that ASOM patients were mainly concentrated in children under 14 years old in Lanzhou. This can be attributed to the anatomical features of the middle ear in children are smaller and shorter than that in adults, and more horizontally aligned Eustachian tubes, and frequent upper respiratory tract infections [13]. The upper respiratory tract plays an important role in filtering and regulating inhaled air. Air is inhaled through the mouth and nose and connected through the nasopharynx. The nasopharynx is connected to the middle ear through the eustachian tube located at the back of the nasopharynx. This direct connection with the middle ear makes the connection between air and the middle ear. As a continuation of the upper respiratory tract, the middle ear mucosa is closely related to respiratory diseases, so the occurrence of ASOM is directly related to respiratory diseases. Many studies have showed that with the aggravation of air pollution, the probability of children suffering from respiratory diseases also increases [17]. Furthermore, air pollution can lead to changes in children's immune function and more prone to inflammation.
Meanwhile, we also found that there were more male ASOM patients than female ASOM patients, and male patients were 1.16 times as many as female patients, this is consistent with the literature report [15]. Which may be related to the higher incidence of acute respiratory infections in male patients, or to the sex ratio of infants born in China. Further large-scale epidemiological studies are needed to determine what factors contribute to this difference. The incidence of respiratory tract infection in children mentioned in some literatures is higher in males than in females.
Seasonal distribution showed that the daily number of ASOM visits in summer was significantly less than that in other three seasons, which is consistent with other findings that OM is more common in winter than in summer [16].The reason may be that there is more rain in summer, nasal mucosa is wet and eustachian tube cilia are moving well, which cause the incidence of ASOM is relatively low. We also found the largest number of ASOM visits occur in December of each year. One important reason is the cold weather in winter, a large number of coal combustion heating and heavy industry operation lead to more serious air pollution, and PM 2.5 , PM 10 , CO, NO 2 , SO 2 are higher than other three seasons, bad climate conditions and human susceptibility body, air pollutants with bacteria and (or) viruses will speed up the invasion of the human body. Moreover, the weather is dry in winter, the respiratory mucosa, especially the eustachian tube and nasal mucosa loses water, and the human resistance decreases in winter, air pollutants such as particulate matter PM2.5 and PM10 will adhere to the respiratory mucosa, they will act as allergens to increase the risk of respiratory infection, and the number of ASOM patients is also higher than other seasons [17]. In addition, people mainly live indoors in winter, creating favorable conditions for the spread of the virus, while influenza mainly spreads through the air. Therefore, low winter temperature and upper respiratory tract infection are risk factors for ASOM.
We can also learned that the lowest number of patients in August, followed by February. On the one hand, the air quality from June to August is better. On the other hand, the students have summer vacation in August. February is the next lowest month for ASOM visits, which is contrary to what we think is the visits increase in winter. We believe that this result is in line with the characteristics of big cities in China. February is usually the Chinese Lunar New Year and the winter vacation for students.
A large number of people who usually work in big cities return home for the New Year or go out for tourism, the number of people in the city and the total number of hospital visits have been reduced, which makes the high incidence month of ASOM become a relative low throughout the year.

Ambient Air Pollution And ASOM
The current research reports that there is a significant statistical correlation between the improvement of air quality and the reduction of ear infection rate. Deng et al. [18] reported that the incidence of OM in urban was higher than that in rural, and urban pollution was more serious than that in rural areas. Heinrich et al think that tobacco exposure and air pollution are both important environmental risk factors for OM, while cdceres and others think that air pollution and low socioeconomic status are more likely to be risk factors for AOM than parents smoking [19]. Although prior studies have demonstrated associations between high ambient air pollutants and OM, these have been small, short-term studies of population cohorts outside the Chinese, and there is no report on the impact of air pollution on the incidence of ASOM and which pollutants can cause or aggravate ASOM. This study We found PM 2.5 , PM 10 Table 5 only showed lag 1-7 data. The multi-day were selected for 1-14 days (Lags01-014, L01-L014). When running the model, we also consider the data of each factor after 14 days, but the correlation is very small. Therefore, the lag results of more than 14 days are not included in Table 6. As for single lag effects, there was a positive correlation between PM 2.5 , PM 10  We also found that the daily number of ASOM visits were negatively correlated with temperature.
That means the lower the temperature, the higher the incidence of ASOM. The common pathogens causing ASOM include Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis, respiratory syncytial virus and rhinovirus. Numminen et al. found that dry and cold season is more conducive to the spread of Streptococcus pneumoniae, so the incidence of AOM increases at low temperature [24]. Old thorium studies such as Kim also found that there was a significant positive correlation between Streptococcus pneumoniae infection and temperature below 24℃, and the incidence of respiratory syncytial virus infection was higher in winter than in other seasons [25]. In addition, we also found that the incidence of ASOM had a positive correlation with ATM, the reason may be that the Eustachian tube also allows air exchange and pressure balance. The shape of exposure-response correlation reflects the potential health effects of atmospheric pollutants and meteorological factors on ASOM. In this study, In this study, we found different exposure-response curves of air pollutants and meteorological factors to the daily number of ASOM visits. At present, no scholar has studied the exposure-reaction relationship between air pollutants and meteorological factors to ASOM, which needs further research to prove.

Biological Mechanism Of ASOM
Although the specific mechanism of air pollutants causing ASOM is unclear, animal studies have shown that air pollutants can inhibit mucociliary clearance of respiratory epithelium, which may have similar effects due to the similar histology of middle ear mucosa. On the one hand, air pollutants can destroy the nasal mucosal epithelial barrier, increase the contact opportunity and time between pollutants and upper respiratory mucosa, and induce and aggravate the occurrence of ASOM. For example, SO2 stimulates mucus secretion in the proximal exposed part of the middle ear and impairs ciliary body function in the distal exposed part. On the other hand, air pollutants are small in volume, easy to combine with mucosal water to produce various products, enhance the permeability of respiratory epithelium, and improve the sensitivity and contact opportunity of mucosal epithelial cells with allergens. In addition, contaminants adhere to the cell surface, resulting in disordered cell arrangement and the destruction of the structure of ciliated cells. It can reduce the activity of mucosal epithelial cells, induce the synthesis and release of inflammatory cytokines, induce apoptosis through oxidative stress, and destroy the barrier function of mucosal epithelial cells. The incidence of allergic rhinitis increases, which indirectly affects the incidence of ASOM. On the other hand, it has been reported that air pollutants can increase the incidence of rhinitis and sinusitis [26]. Sinusitis is one of the main causes of otitis media. The retrograde transport of purulent secretions by nasal cilia leads to eustachian tube edema and nasopharyngeal infection into the middle ear, which leads to the irremovable secretions of the middle ear cavity. It is the mechanism of otitis media caused by sinusitis. Therefore, the increase of air pollutant concentration can indirectly increase the incidence of ASOM. Mostly, exposure to air pollution can lead to immune and inflammatory dysfunction and increase the risk of allergic rhinitis [23]. Andrianifahanana et al. [27] showed that the incidence of OM in patients with allergic rhinitis was high, because nasal lesions such as inflammation and allergy were the main causes of dysfunction of eustachian tube, and dysventilation and drainage of eustachian tube were also directly related to the occurrence of ASOM. However, not only in low income countries, wood heaters or fire places are also common in some developed countries. Studies have shown that the results of increased exposure to wood smoke and OM risk are consistent with the toxicological effects of wood smoke on respiratory epithelial cells [28,29]. In the experimental study, the composition of wood smoke increased oxidative stress in epithelial cells [30]. Recent experimental studies have shown that human respiratory cells exposed to particulate matter from beech sawdust smoke have cytotoxic and genotoxic effects comparable to those produced by diesel engines [31]. As true upper respiratory mucosa, middle ear mucosa has similar mechanism of causing ASOM.

Conclusions
This study provided evidence of the adverse effect of ambient air pollution on ASOM in Northwestern China. We analyzed the distribution characteristics of air pollutants and meteorological factors in Lanzhou from 2014 to 2016, and the correlation between air pollutants and meteorological factors.
We also examined the lag effects of air pollutants, meteorological factors and the possible differences of air pollutants, meteorological factors on the daily number of ASOM visits that were consulting the hospital for treatment for ASOM over a period of 3 years. The results showed that the daily number of ASOM visits had obvious seasonal differences. The air pollutants positively correlated with the daily number of ASOM visits, and had lag effects on the daily number of ASOM visits, temperature is negatively correlated with the daily number of ASOM visits. RR in the number of ASOM visits increased with air pollution level. There are also some limitations to our study. The data at the moment are so sparse that we are not in a position to provide any far-reaching conclusion yet.
Further experimental studies are needed to prove. With the continuous rapid urbanization process, more people are becoming exposed to high levels of air pollution. Environmental control and public health strategies should be enforced by the health service policy makers to address this increasingly challenging problem. During haze events, both the health care provider and the public should be Presents the Single-lag exposure-response relationships for air pollutants and meteorological factors with hospital visits for ASOM , the solid line represents logarithmic relative risk (log RR) and the dashed line represents 95% confidence interval (CI).
BAR=ATM Figure 5 Presents the cumulative lag exposure-response relationships for air pollutants and meteorological factors with outpatient visits for ASOM inLanzhou, China (2014-2016), the solid line represents logarithmic relative risk (log RR) and the dashed line represents 95% confidence interval (CI). BAR(=ATM) 3.5 Relative risk

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.