Introduction

Meteorological factors including temperature, humidity, wind speed, air pressure, and many other factors are closely related to human health (Liang and Gong 2017). In recent years, severe climate change such as global weather events with extreme surface temperatures has become more frequent and intense, taking the experience of the 2003 European summer heat wave as an example, which directly caused 44,000 additional deaths (Tobias et al. 2012). Severe meteorological (climate) changes are expected to continue to dominate the world in the coming decades (Peters et al. 2022). With the massive use of fossil fuels and modern industrial development leading to unstable climate change, there is growing concern about the direct impact of a range of extreme weather-related events on public health, such as temperature, humidity, or rainfall, which have been identified as risk factors for the spread of some common clinical diseases such as malaria (Ninphanomchai et al. 2014), childhood diarrhea (Bradatan et al. 2020), and COVID-19 (Zhang et al. 2020). In contrast, the risk of developing eye surface problems such as conjunctivitis, keratitis, and dry eye disease, which are directly exposed to air (Zhang et al. 2021; Mandell et al. 2020), may be influenced by environmental meteorological factors such as temperature. The ocular surface can be affected by direct physical and chemical stimuli caused by meteorological factors, such as temperature (Craig et al. 2000), humidity (Cekic et al. 2002), and ultraviolet radiation (Zuclich and Connolly 1976), which contribute to the induction of oxidative stress and inflammation in the ocular surface and ultimately lead to imbalance and dysfunction of the intraocular environment and corresponding diseases. But the number of available relevant studies is very limited, and further detailed research is urgently needed.

Conjunctivitis is an inflammatory disease of the conjunctival tissues and is one of the most common ophthalmic diseases, which manifests clinically with increased discharge, conjunctival congestion, photophobia, and itchy sensations. The conjunctiva is directly exposed to the air and is therefore susceptible to immune inflammatory reactions due to exposure to environmental particles or allergens. Conjunctivitis significantly affects the patient physically and psychologically with a significant loss of quality of life (Zhang et al. 2017, 2021). In the USA, outpatient visits for conjunctivitis account for 1/3 of annual visits for ophthalmology-related diseases, while data from the Chinese Center for Disease Control and Prevention suggest that conjunctivitis cases in China are also increasing annually, with allergic conjunctivitis patients alone increasing by 0.3% in 2020 compared to the previous year, reaching 295 million (Azari and Barney 2013). According to an emergency ophthalmology study in the USA, the economic burden of eye diseases in emergency room visits is estimated at $2 billion per year, with conjunctivitis alone accounting for 28% of the total (Channa et al. 2016). Therefore, it has become a public health problem that cannot be ignored (epidemiology of eye-related emergency department visits). As a multicausal disease, numerous risk factors for conjunctivitis, such as pollen (Sheng et al. 2022), air pollutants (Nucci et al. 2017), and dust (Chien et al. 2014), have been shown to be associated with the risk of conjunctivitis. The impact of changes in ambient environmental conditions including temperature, humidity, and atmospheric pressure on conjunctivitis has received insufficient attention, but a series of recent studies have confirmed that multiple meteorological factors may be associated with the risk of conjunctivitis while showing potential lagged effects and geographic variation (Zhang et al. 2021; Patel et al. 2021), but the number of relevant studies remains limited. In addition, except for a few types such as allergic conjunctivitis, most other different types of conjunctivitis have not been studied in detail.

Urumqi is one of the most economically developed large cities in Northwest China and the capital of the Xinjiang Uyghur Autonomous Region (hereinafter referred to as “Xinjiang”), with a vast land area but little actual habitable land. Urumqi is the world’s most inland city, farthest from the sea and coastline (2500 km), also known as the “capital of Asia,” the special geographical location of Urumqi makes it a suitable place to study the effects of continental climate on health. In addition, local long-term and periodic changes in meteorological factors with obvious differences provide an excellent place to study the impact of meteorological factors (especially extreme meteorological factors) on health. It has unique long-term climate characteristics, such as dry and windy, and frequent extreme weather events, such as sandstorms. At the same time, it shows the typical mid-temperate continental arid climate with short spring and autumn, long winter and summer, significant temperature difference between day and night, dramatic changes in winter and summer, long and cold winter, and less annual precipitation, which combined with other factors such as severe air pollution may lead to weather changes that significantly affect human health, such as coronavirus disease 2019 (COVID-19) (Yang et al. 2021); hand, foot, and mouth disease (Ren et al. 2022); and acute aortic coarctation (Shi et al. 2017). Considering the increasing number of studies that have focused on the relationship between conjunctivitis and meteorological factors, especially extreme weather (Hong et al. 2016), there are still few studies that have examined the association between meteorological factors and conjunctivitis in Urumqi, the farthest large inland city from the ocean with highly representative climate characteristics, not to mention examining the lagged effects of long-term or extreme meteorological factors and the variations in the effects on different types of conjunctivitis.

Therefore, we used a time-series study method and the matched lag model to explore the relationship between local meteorological factors including temperature, humidity, air pressure, wind speed, and conjunctivitis outpatient visits in Urumqi, the major northwestern city with specific climatic and geographical characteristics. The lag patterns of the four major meteorological factors in conjunctivitis outpatient clinics in Urumqi were systematically investigated, and subgroup analyses were conducted according to age, gender, four seasons, cold and warm weather to explore further lag patterns of meteorological factors on different types of conjunctivitis outpatient visits and further analyzed the effects of different conjunctivitis types and extreme meteorological factors.

Methods and materials

Study site and climatic characteristics

Our study was conducted in Urumqi, Xinjiang, China. Urumqi City is in the central part of Xinjiang Province in northwest China, in the hinterland of the Asia-Europe continent, at the northern foot of the Northern Tianshan Mountains, and in the southern part of the Junggar Basin. It is an important central city in Northwest China and an international trade center for Central Asia and West Asia, with a total area of 13,800 km2 and a resident population of 4.07 million by 2021. It is the political, economic, cultural, scientific, educational, and transportation center of Xinjiang.

With its undulating terrain and vast mountainous terrain, Urumqi is characterized by an aggregated population distribution. As the most inland city in the world and the farthest from the ocean and coastline (2500 km), Urumqi is known as the “Capital of Asia” and therefore exhibits a characteristic mid-temperate continental arid climate, characterized by an arid climate throughout the year and scarce precipitation that increases vertically with altitude. The seasons are unevenly distributed, which means that the spring and autumn seasons are short, and the winter and summer seasons are long, especially the winter season is cold and extremely long, reaching an average of 5 months. There is a large temperature difference between day and night, and the cold and heat change drastically. The hottest month is July and August, with a mean temperature of 25.7 ℃; the coldest month is January, with a mean temperature of − 15.2 ℃.

Study population

The electronic medical record system outpatient data from January 1, 2013, to December 31, 2020, were obtained from the Ophthalmology Department of the First Affiliated Hospital of Xinjiang Medical University (Urumqi, Xinjiang, China), which is the largest as well as the most modern ophthalmology clinic in Xinjiang Province and the most preferred medical institution for patients with ophthalmic diseases in Xinjiang and the surrounding areas including Urumqi.

Basic information such as gender, age, and residence address (including postal code) were extracted first. Patients from emergency clinics, re-visits, and those who had lived in Urumqi for less than 6 months were excluded, and only information from first-time patients was retained. The ophthalmology section of the International Classification of Diseases Standard codes (ICD-10) (including H10.901, H10.902, H10.301, H10.402, H10.801, H10.101, H10.102, H10.201) was used for the classification and diagnosis of conjunctivitis. The detailed geographical location of the hospitals is shown in Fig. 1 and Supplementary Material B.

Fig. 1
figure 1

Location of environmental monitoring stations and hospital in Urumqi, Xinjiang. The location of each air quality monitor is marked with a black triangle; the location of the hospital is marked with an aquamarine cross; the home address of each recruited patient is marked with an orange-red dot. FAHXMU, First Affiliated Hospital of Xinjiang Medical University (Urumqi, Xinjiang, China); 1, Urumqi Toll House; 2, Urumqi Thirty-first High School; 3, Urumqi Xin Shi Da Hot Spring Campus; 4, Urumqi Hongguang Mountain Area; 5, Urumqi Great Green Valley; 6, Urumqi Xinjiang Academy of Agricultural Sciences Farm; 7, Urumqi Monitoring Station; 8, Urumqi Railway Bureau; 9, Urumqi Middong District Environmental Protection Bureau; 10, Urumqi Darbancheng District Environmental Protection Bureau; Urumqi Training Base Protection Bureau; 11, Urumqi Training Base

Meteorological and outdoor air pollutant information

Meteorological data for daily maximum temperature (℃), daily minimum temperature (℃), daily mean temperature (℃), daily relative humidity (%), daily average wind speed (m/s), and atmospheric pressure (hPa) from January 1, 2013, to December 31, 2020, were obtained from the China Meteorological Data Sharing Service (http://data.cma.cn). In addition, considering the few previous studies suggesting the effect of air pollution on conjunctivitis, the main air pollution data, including PM2.5 (daily 24-h average), PM10 (daily 24-h average), CO (daily 24-h average), NO2 (daily 24-h average), and SO2 (daily 24-h average) and O3 (maximum daily 8-h average), were extracted and analyzed from 11 local standard urban background stationary air quality monitors, and the detailed geographical location and distribution of the monitoring sites are shown in Fig. 1 and Supplementary Material B.

Statistical analysis and model construction

Spearman correlation coefficients were used to assess the correlation between meteorological factors and air pollutants. the R package “seasons” was used to set alternative values for the control groups in the model. “Day of the week (DOW)” was used as an indicator variable to reconcile long-term trends, seasonal pattern effects, and day of the week effects. To assess potential lagged effects of meteorological factors (including extreme weather), we included lags of up to 7 days in the model for cumulative and noncumulative exposures, given the latency period of conjunctivitis of up to 5–6 days. Cumulative effects were defined as mean concentration effects for lags 0–n days, and noncumulative effects were defined as exposure effects for lag 0 to lag n days. The optimal lag for each lag model was obtained based on the maximum value of relative risk (RR) and the minimum value of p.

Daily outpatient visits for conjunctivitis are considered rare events that approximate a quasi-Poisson distribution, so we used a quasi-Poisson generalized linear regression model with a distributed lagged nonlinear model (DLNM) to fit the effects of meteorological factors (including extreme weather) on outpatient visits for conjunctivitis, and the effects of confounding variables, including the major six air pollutants, were controlled for by a natural cubic spline curve (ns) with three degrees of freedom (df) for smoothing control. df was chosen as the minimum of the sum of the absolute values of the partial autocorrelation function (PACF) based on the residuals of the underlying model, consistent with the residual independence principle.

The conjunctivitis clinic data were processed by firstly screening out conjunctivitis diagnosis records without missing information (n = 64,613) and secondly screening out patients living in the main area of Urumqi city with an average distance of less than 20 km from the nearest monitoring station (n = 59,731) according to the Baidu Map 0.2 package in the R software. Air pollution factors were considered as covariates in the model, and the final model is shown below.

$${Y}_{t}\sim \mathrm{quasiPoisson }\;({\mu }_{t})$$
$$\begin{array}{l}\mathrm{Log}\left[E\left({Y}_{t}\right)\right]=\beta \times {Z}_{t}+\mathrm{factor }\left(\mathrm{DOW}\right)+\mathrm{ns }\left(\mathrm{time},\mathrm{ df}=7/\mathrm{year}\right)+\mathrm{ns }\left({\mathrm{NO}}_{2},\mathrm{ df}=3\right)\\ +\mathrm{ns }\left({\mathrm{O}}_{3},\mathrm{ df}=3\right)+\mathrm{ns }\left({\mathrm{SO}}_{2},\mathrm{ df}=3\right)+\mathrm{ns }\left(\mathrm{CO},\mathrm{ df}=3\right)+\mathrm{ns }\left({\mathrm{PM}}_{2.5},\mathrm{ df}=3\right)\\ \begin{array}{l}+\mathrm{ns }\left({\mathrm{PM}}_{10},\mathrm{ df}=3\right)+\mathrm{ns }\left(\mathrm{T},\mathrm{ df}=3\right)+\mathrm{ns }\left(\mathrm{RH},\mathrm{ df}=3 \right)+\mathrm{ns }\left(\mathrm{AP},\mathrm{ df}=3\right)\\ +\mathrm{ns }\left(\mathrm{WS},\mathrm{ df}=3\right)+\mathrm{intercept}\end{array}\end{array}$$

E(Yt) denotes the estimated number of outpatient visits for patients with conjunctivitis on day t, and Zt is the level of certain meteorological factors on day t, where T, RH, AP, and WS represent mean air temperature, relative humidity, atmospheric pressure, and wind speed, respectively; β is the exposure coefficient; ns is the natural cubic spline function; df is the degree of freedom, and DOW denotes the day of the week. We report two types of results for the effects: On the one hand, we report RRs and 95% confidence intervals (CIs) for continuous exposure (per 10 units increase compared to the average background exposure level), and on the other hand, taking into account the unique climatic characteristics of the region, we report the effects of climate extremes according to categorical exposure (extremely high/low vs. level of average (P50)) models by setting “extremely low levels” of 0–10% and “extremely high levels” of 90–100% for each meteorological factor.

To investigate the effect of extreme meteorological factors on conjunctivitis outpatient visits, we compared 1% (P1), 5% (P5), 10% (P10), and 90% (P90), 95% (P95), and 99% (P99) of each meteorological factor with the average level (50%) to obtain the effect of extremely low levels as well as extremely high levels of exposure on conjunctivitis outpatient visits.

Finally, to explore differences in the response of different types of conjunctivitis to air pollution, we classified conjunctivitis as acute, chronic, infectious (mainly including bacteria, viruses, chlamydia, etc.), allergic, follicular, vesicular, and nonspecific. To further explore potential modifying factors, we performed subgroup analyses for sex (male, female) and age (0–1, 2–5, 6–18, 19–64, and ≥ 65 years) to further explore potential modifying factors. Considering the great climatic variation of local seasons, our local seasons were divided into spring (10–22 °C April–May), summer (> 22 °C June–August), autumn (10–22 °C September–October), and winter (< 10 °C November–March) according to international climate criteria (mean temperature for 5 consecutive days). Statistical analyses were performed in R software version 4.0.4 (February 15, 2021) using the main packages including “seasons,” “dlnm,” and “splines.” Statistical tests were all two-sided, and P values less than 0.05 were considered statistically significant. The study was conducted in accordance with the Declaration of Helsinki, and the Medical Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University approved and supervised the study in its entirety.

Results

The demographic data of the patients included in this study are shown in Table 1. Females accounted for nearly 59% of the 59,731 patients diagnosed, approximately 60% (35,311) of the study participants were young adults aged 19–64 years, 12.09% (7220) of newborns with conjunctivitis were diagnosed in 0–1, and older adults over 65 years of age (4027) were the least diagnosed group (6.7%). Diagnosis of conjunctivitis was higher in the warm season compared to the cold season, with the highest in the summer (July–August) and the lowest in the winter (November–February). Table 1 also shows the concentrations of the six main air pollutants and the four meteorological factors (relative humidity, mean temperature, air pressure, and wind speed) in Urumqi for 2013–2020. The average concentrations of CO (24 h), NO2 (24 h), SO2 (24 h), PM2.5 (24 h), PM10 (24 h), and ozone (8 h) are 1.2 mg/m3, 48.5, 16, 64.5, 120.2, and 68.2 μg/m3, respectively. The average values of meteorological data: mean temperature (− 26 to 35.1 °C), relative humidity (6–100%), air pressure (842–934 hPa), and wind speed (0–14.8 m/s) were 8.4 °C, 55.5%, 912.7 hPa, and 1.9 m/s. The remaining descriptive statistics (standard deviation (SD), minimum, 25th and 75th percentiles, median, and maximum) and details of the period variation characteristics are shown in Fig. 2 and Supplementary Material A, Figure S1.

Table 1 Characteristics of outpatient for conjunctivitis in the First Affiliated Hospital of Xinjiang Medical University (January 1, 2013 to December 31, 2020)
Fig. 2
figure 2

Distribution of exposure to four meteorological factors (mean temperature, relative humidity, atmospheric pressure, wind speed) and outpatient visits for conjunctivitis over time in Urumqi, Xinjiang, January 1, 2013–December 31, 2020

The results of the correlation analysis suggested that there were different degrees of correlation between different types of meteorological factors and air pollutants, with a strong correlation between mean temperature and relative humidity (r = 0.81, p-value < 0.0001) and the correlations between O3 and mean temperature (r = 0.81, p-value < 0.0001) as well as relative humidity (r =  − 0.72, p-value < 0.0001) were also strong (Fig. 3).

Fig. 3
figure 3

Spearman correlation between different air pollutants and meteorological factors. T, temperature; RH, relative humidity; WS, wind speed; AP, atmospheric pressure

Results of the single-meteorological factor variable model showed that only mean temperature and atmospheric pressure factors were associated with outpatient visits for conjunctivitis, with per 10-unit increase in ambient temperature at lag days 1 (RR = 1.063) and 2 (RR = 1.056), and cumulative lags 0–2 (RR = 1.080), 0–3 (RR = 1.092), and 0–4 (RR = 1.079) were statistically significantly associated with increased conjunctivitis outpatient visits. In contrast to this effect, the air pressure factor showed a protective effect, specifically per 10 unit increase in atmospheric pressure was statistically significantly associated with decreased conjunctivitis visits at the lag 7 day (RR = 0.967) and cumulative lags 0–2 days (RR = 0.944), 0–3 days (RR = 0.947), and 0–7 days (RR = 0.921). It failed to observe statistically significant results for relative humidity as well as wind speed under the single meteorological factor model (Table 2).

Table 2 The relative risks (RRs) of per 10 units increased in meteorological variables on conjunctivitis outpatient visits at various lag days: single-meteorological variable model

The results of the multimeteorological factor model were like those of the single-meteorological factor, with mean temperature and atmospheric pressure similarly showing significant correlations with increased (RR = 1.058–1.114) and decreased (RR = 0.920–0.967) conjunctivitis outpatient visits, respectively. However, each 10-unit increase in relative humidity exhibited significant correlations with increased conjunctivitis outpatient visits in the multimeteorological factor model (RR = 1.019–1.025) (Table 3). The dose–response curves for these four meteorological factors and the risk of conjunctivitis outpatient visits are displayed in Fig. 4 and Supplementary Material A, Figure S2-210.

Table 3 Correlation of per 10 units increased in air pollutants and outpatients for conjunctivitis: multi-meteorological variables model
Fig. 4
figure 4

Overall exposure–response association for meteorological factors and outpatient visits for conjunctivitis: single-variable model

The results of the subgroup analysis showed gender differences in the effect of mean temperature on conjunctivitis, specifically the cumulative lagged effect of mean temperature was significantly associated with decreased outpatient visits in male patients with conjunctivitis (RR = 0.514, 95% CI = 0.265–0.997), while it was not detected in female patients, but the single-day lagged effect of mean temperature was significantly associated with increased outpatient visits in female patients with conjunctivitis (RR = 1.082, 95% CI = 1.009–1.161). Age subgroup results suggested the protective effect of per 10-unit increase in mean temperature for newborns aged 0–1 year (RR = 0.908 for single-day lag, 0.232 for cumulative lag), while relative humidity (RR = 1.014), as well as wind speed (RR = 2.077), was significantly correlated with increased outpatient visits for patients with conjunctivitis aged 0–1 year. Increased atmospheric pressure was significantly associated with an increased risk of conjunctivitis in the underage population aged 6–18 years (RR = 2.467 for single-day lag, 2.746 for cumulative lag). The conjunctivitis population aged 16–64 years was only significantly correlated with mean temperature, while no statistically significant results were found in the older conjunctivitis population aged > 64 years. Subgroup analysis of the seasonal factors shows that only relative humidity effects exhibited differences between the warm and cold seasons (RR = 1.008 for the warm season). Further grouping of the four seasons suggests that the effects of relative humidity, atmospheric pressure, and wind speed exhibit seasonal differences, with the most significant effects mainly in summer and winter (Table 4).

Table 4 Correlation of per 10 units increased in meteorological variables and outpatients for conjunctivitis and effect modification through stratified by patients’ characteristics

The relationship between different types of conjunctivitis and different meteorological factors varies. Specifically, our results suggest that acute conjunctivitis is only associated with atmospheric pressure (RR = 1.147). However, in chronic conjunctivitis, the increase in atmospheric pressure showed exactly the opposite effect (RR = 0.728), like the relative humidity (RR = 0.728). Allergic conjunctivitis showed a significant correlation with an increase in mean temperature and relative humidity (RR = 1.216 and 1.553), but there was a significant correlation between increased wind speed and decreased outpatient visits for allergic conjunctivitis (RR = 0.220). In addition, follicular conjunctivitis showed significant correlations with increases in relative humidity and atmospheric pressure (RR = 0.855 and 4.037), and nonspecific conjunctivitis showed significant correlations with increases in mean temperature (RR = 1.076) (Table 4).

The effect of extreme weather is demonstrated in Table 5, where the results show that low levels of atmospheric pressure (for P1 RR = 1.092, for P5 RR = 1.055, for P10 RR = 1.040) and relative humidity (for P1 RR = 1.123, for P5 RR = 1.084, for P10 RR = 1.067) increase the risk of conjunctivitis outpatient visits and the effect increased with the decreasing levels. Extreme high levels of mean temperature were statistically correlated with increased outpatient visits for conjunctivitis and this effect increased progressively with increasing mean temperature (for P90 RR = 1.112, for P95 RR = 1.135, and for P99 RR = 1.164). Extreme wind speed demonstrated the protective effect on outpatient visits for conjunctivitis, which increased with decreasing wind speed levels under very low levels, and the protective effect was strongest at extremely low wind speeds (for P1 RR = 0.954). The results were similar under single meteorological factor as well as multimeteorological factor models.

Table 5 The relative risks (RRs) under extreme weather and outpatients for conjunctivitis: single-meteorological variable model. Extremely low levels P1, P5, P10 (1%, 5%, 10%); extremely high levels P90, P95, P99 (90%, 95%, 99%); mean level P50 (50%)

In addition, we also investigated the relationship between six major air pollutants and conjunctivitis outpatient visits in Urumqi from January 1, 2013, to December 31, 2020. The results of single-pollutant and multipollutant models suggested that PM2.5, O3, and SO2 did not show significant correlations with conjunctivitis outpatient visits, and PM10 (RR = 1.004–1.005), NO2 (RR = 1.019–1.046) and CO (RR = 0.981) showed weak correlations with conjunctivitis outpatient visits in Urumqi. The results of single-lag and cumulative-lags are shown in Supplementary Material A, Table S1-2. Overall, these results suggest that there may be little significant substantial correlation between air pollution and local conjunctivitis outpatient visits in Urumqi.

Discussion

Conjunctivitis can affect people of all ages and genders, imposing a significant economic burden and severely diminishing the ability of patients to learn and work effectively. Approximately 15–40% of the population in developed countries experience a reduced quality of life from conjunctivitis (Pitt et al. 2004), for example, 3–6 million people suffer from conjunctivitis each year in the USA, with acute conjunctivitis comprising the largest number (Ramirez et al. 2017; Azari and Barney 2013). A report from Hangzhou, China, indicates that the average cost of examination and treatment of conjunctivitis is estimated to total about $30 per case (Fu et al. 2017), and similar conditions exist in other developing countries.

Urumqi is in the hinterland of the Eurasian continent and is the world’s farthest city from the ocean, nicknamed the “Capital of Asia,” and is an important gateway for China's economic development to the west as the second bridgehead of the Asia–Europe Continental Bridge in western China. Until 2021, it has a permanent population of 4.07 million and is the main transportation route and local political, economic, cultural, technological, educational, and transportation center in Northwest China. In recent decades, the local economy and society have developed rapidly, so more and more people have begun to pay attention to the impact of local environmental factors such as extreme meteorological factors on health problems. As the most inland city in the world and the farthest from the ocean and coastline (2500 km), Urumqi located in the hinterland of Eurasia, the northern foot of the Tianshan Mountains, southern Junggar Basin, covers an area of 12,000 km2 and has a vast mountainous area that blocks ocean water vapor from reaching the city, so the climate is arid throughout the year, showing typical characteristics of a temperate continental arid climate, with the hottest months being July and August (mean temperature 25.7 °C) and the coldest month being January (mean temperature − 15.2 °C) (Hu et al. 2018). Spring and autumn are short in duration while winter and summer are long. Due to drought and other factors resulting in less surface cover vegetation, the Urumqi region is highly susceptible to windy weather and the formation of dust storms in the downstream region by affecting the exposed surface dust source sites, which not only causes meteorological disasters for agricultural production and transportation in the Xinjiang region but also has a significant impact on the day-to-day life of residents (Cui et al. 2020). However, until now there has been no study conducted in this largest oasis city in China, which is undergoing rapid urbanization, to investigate the influence of meteorological factors and local ocular surface diseases. At the same time, although most of the residents in Xinjiang are concentrated in this area, due to the changeable climatic conditions, the suitable habitats are very few, and the local population distribution shows the characteristics of regional aggregation and large-scale unmanned areas. The poor mobility of the population and the distribution characteristics of aggregation have greatly improved the reliability and validity of local research on environmental related health problems, especially providing a good example for studying the impact of extreme meteorological factors on health.

We conducted the first time-series analysis design with a large sample size in Xinjiang, covering 59,731 conjunctivitis cases in Urumqi from 2013 to 2020, and revealed the association of four major meteorological factors with conjunctivitis outpatient visits in the region by analyzing the average level of exposure per 10-unit increase as well as the extreme level of exposure, with a detailed analysis of the respective dose–response relationships. Our results suggest that increased mean temperature in the Urumqi region is significantly associated with an increase in local conjunctivitis clinic visits and that the effect is strongest at extremely high temperatures, specifically that the risk of conjunctivitis outpatient visits is 16.4% higher at mean temperatures above 28.7 °C (P99) than at mean temperatures (10.7 °C); it was also increased at extremely low-temperature levels (− 10.5 °C) (RR = 1.057), as confirmed by our dose–response curves for mean temperature and risk of conjunctivitis outpatient visits, with a “U”-shaped curve suggesting that both extremely high and extremely low temperatures increase the risk of conjunctivitis. In addition, we also identified the lagged effect of elevated temperature on the risk of conjunctivitis. Specifically, the single meteorological factor model revealed that per 10-unit increase in temperature level was significantly associated with an increased risk of conjunctivitis at single-day lags of 1 and 2 days, and cumulative lags of 0–2, 0–3, and 0–4 days, which suggested that elevated mean temperature may impact on the risk of conjunctivitis 1–4 days after exposure.

For relative humidity, our results suggest that low levels of relative humidity increase the risk of outpatient visits for conjunctivitis, with a 12.3% higher risk of outpatient visits for conjunctivitis at the extremely low level of 13% relative humidity (P1) compared with the average level (54%), and this effect increases progressively with decreasing relative humidity (Table 5). However, this effect was not detected in the single meteorological factor model, and in the multiple meteorological factor model, only a slight effect was reflected (RR = 1.021–1.025).

It is worth mentioning that we failed to detect statistically significant correlation results between wind speed and conjunctivitis outpatient visits in single as well as multimeteorological factor models, but the results of extreme levels of exposure suggest that wind speed is a protective factor, and at low levels of wind speed exposure (P1–P10), the risk of conjunctivitis outpatient visits decreased significantly with decreasing wind speed (RR = 0.954–0.975). On the one hand, this may be closely related to the severe air pollution in Urumqi (Ren et al. 2022). On the other hand, the peculiar local climate of prolonged high temperature and low humidity typical of aridity increases surface evaporation, which leads to increased concentrations of irritants such as pollen in the air and dryness of the ocular surface, accelerating the development of conjunctivitis (Mizuno et al. 2021).

The results of the subgroup analysis showed that only temperature showed gender differences in the risk of outpatient visits for conjunctivitis, and previous studies suggest that gender differences in ocular surface disease can be explained by sex hormone-induced intraocular inflammatory responses, as supported by the study of Truong et al. (2014). Whereas for newborns aged 0–1 year, per 10-unit increase in temperature as a protective factor reduced the risk of conjunctivitis, relative humidity, as well as wind speed, increased the risk of outpatient visits for neonatal conjunctivitis, with wind speed presenting the strongest effect (RR = 2.077). For minors aged 6–18 years, a significant correlation was found only between air pressure and increased conjunctivitis risks (RR = 2.467). In contrast, for patients in the 18–64 age group, only temperature was detected as the statistically significant risk factor. Age differences may be associated with low metabolic and immune capacity at different ages such as in infants and the elderly (Chen et al. 2021). In addition, previous studies have suggested that the onset of conjunctivitis tends to show seasonal variations, and our results suggest the relative humidity is significantly associated with the risk of conjunctivitis during the warm season, while a further grouping of the four seasons could see relative humidity and air pressure as risk factors for conjunctivitis in summer as well as in winter, while wind speed is a protective factor similar to previous results, which may be related to the fact that elevated temperature and humidity provide a suitable environment for bacteria and allergens to live (Stockmann et al. 2013).

In addition, further typing of the types of conjunctivitis revealed that the effects of different meteorological factors differed. Weather conditions are known to induce and exacerbate allergic diseases such as allergic conjunctivitis, through the interplay of pollen and air pollutants (McMichael et al. 2006). Our findings similarly confirm that the increase in outpatient visits for allergic conjunctivitis in the Urumqi region was statistically significantly correlated with an increase in mean air temperature as well as relative humidity, while increased wind speed exhibited the opposite effect, similar to the results of a time-series study on allergic conjunctivitis published by Jiaxu Hong et al. in 2016, which may be related to the availability of allergens of suitable temperature and humidity conditions as well as airborne routes (Hong et al. 2016). Contrary to the effect of increased atmospheric pressure on acute conjunctivitis versus chronic conjunctivitis, a study based on a Taiwanese population published by Chun-Chi Chiang et al. in 2012 suggested that chronic conjunctivitis is characterized by two peaks in incidence per year compared to acute conjunctivitis (Chiang et al. 2012), especially during the warmer months, which may indicate seasonal and environmental variations associated with the etiology of the chronic course, thus producing different responses to different meteorological factors.

There is no consensus on the exact mechanism of action of meteorological factors on the development of conjunctivitis. However, in recent years, with the increase in climate change and the frequency of extreme weather times, more and more attention has been paid to the relationship between meteorological factors and disease, and many studies have revealed that meteorological factors such as temperature and air pollutants could exert some influence on human health, especially the extreme weather effects are stronger (Rocque et al. 2021). Air pollutants have been widely recognized as an important risk factor for the development of conjunctivitis, and several previous studies have shown that meteorological factors modulate the local concentrations and effects of air pollutants and indirectly play a role (Fu et al. 2017; Lu et al. 2019). A study published by Iny Jhun et al. in 2014 suggested that the association between ozone and mortality in 97 US cities becomes stronger in hotter and colder environments (Jhun et al. 2014). In addition, high wind speed can promote the dispersion of outdoor particulate matter (including pollen and mold spores) and thus reduce its concentration, and high temperature and solar radiation can increase the effect of photosynthesis on ozone and thus affect the development of allergic diseases (e.g., eczema, allergic conjunctivitis) (Novaes et al. 2010), and our correlation analysis also confirmed this, suggesting a correlation between mean temperature and relative humidity and ozone in Urumqi (r = 0.81 and − 0.72, p-value < 0.001). In addition, inflammatory cell activity may be elevated in high-temperature environments, leading to inflammation in the eye, especially the ocular surface and possibly playing a role in the development of conjunctivitis (Chen et al. 2018). In addition to acting on the inflammatory response, low temperatures can also affect the function of the ocular surface and related diseases by modulating the level of oxidative stress in concert with air pollutants (Duan et al. 2019). Variations in temperature as well as air pressure are not only directly related to the stability of the tear film, affecting the lubricating and protective capacity of the eye (Mandell et al. 2020), but also can further directly induce the development of conjunctivitis through ocular surface dynamics (Abusharha et al. 2016). In addition, changes in temperature, wind speed, air pressure, and humidity can affect not only the concentration, distribution, and composition of particulate matter and gaseous pollutants, but also evaporation from the surface, thereby affecting the concentration and transmission of airborne allergens such as pollen (D’Amato and Cecchi 2008), of which pollen sensitivity is widely considered to be the most common factor in allergic conjunctivitis, and the association may vary with the seasonality of meteorological factors leading to seasonal changes in conjunctivitis outpatient visits (Leonardi et al. 2015).

A growing number of studies in recent years have shown that many human health problems, including ocular surface diseases, may be adversely affected by changes in weather patterns, especially extreme meteorological factors (Lu et al. 2019). The effects of weather factors on neurodegenerative pathologies (Bongioanni et al. 2021), cardiovascular system diseases (Khraishah et al. 2022), and some infectious diseases (Williams et al. 2021) have been extensively studied, but there are still fewer studies on the effects of direct exposure to the environment on ocular surface diseases. In a nationwide, population‐based, cross‐sectional study, Lee et al. found that patients with allergic conjunctivitis had increased incidence in May (spring), September (autumn), and the valley in winter (Lee et al. 2020). However, data from a case-crossover study between the years 2009 and 2014 conducted by Khalaila et al. suggested that patients for conjunctivitis increased during summer and autumn, excluding spring and winter (Khalaila et al. 2021). In addition, a study based on the population in Northeast China, between the years 2014 and 2018, showed that with the increase in temperature and the decrease in humidity, allergic conjunctivitis appears to increase significantly, indicating dry and hot climate is likely to induce allergic conjunctivitis (Lu et al. 2021). It is interesting that long-term exposure to higher mean temperature and relative humidity were associated with an increased risk of acute hemorrhagic conjunctivitis in a study by Zhang et al. (2021). Similarly, in another major study, Seo et al. demonstrated that the number of conjunctivitis outpatients was elevated with higher exposure to temperature and humidity (Seo et al. 2018) Conversely, Sheng et al. published a time-series analysis in 2022, which indicated that there was a negative correlation between temperature and allergic conjunctivitis visits (Sheng et al. 2022). Our results showed that increased mean temperature and extremely low levels of relative humidity in Urumqi were significantly associated with increased local conjunctivitis outpatient visits, while increased atmospheric pressure and extremely low levels of wind speed were significantly associated with decreased local conjunctivitis outpatient visits. In addition, the dose effect of local temperature and conjunctivitis risks showed a “U”-shaped curve, suggesting that extreme levels (very high or very low) of temperature conditions are also risk factors for conjunctivitis. These results suggest that the relationship between meteorological factors and conjunctivitis visits is different in Urumqi than in other regions, which may be related to its unique geographical location and social development pattern (Ren et al. 2022) (Yao et al. 2018).

Our study has several advantages. First, this is the first large-scale time-series analysis study with a long-time span in a large city farthest from the ocean in the world and is the first exploratory investigation of local meteorological factors and the correlation between extreme weather and conjunctivitis clinic visits in Urumqi, which is the largest city in northwest China. Second, our study provides new evidence in this field, confirming for the first time that mean temperature and extremely low levels of relative humidity in Urumqi are risk factors for conjunctivitis, while elevated air pressure and extremely low levels of wind speed are protective factors for conjunctivitis, and suggesting the potential lag effect of mean temperature and air pressure. Third, we further distinguished specific types of conjunctivitis and as comprehensively as possible investigated different meteorological factors and exposure to extreme weather in relation to the outpatient risk of specific types of conjunctivitis and performed subgroup analyses to detect differences in the effects of gender, age, and cold and warm as well as seasonal factors in local conjunctivitis patients in Urumqi.

It is undeniable that our study still has some limitations. First, although the Ophthalmology Department of the First Affiliated Hospital of Xinjiang Medical University, the largest and most modern ophthalmology clinic in the region, which attracts more than 50% of dry eye patients to outpatient clinics, is the preferred medical institution for patients with ophthalmic disease in the surrounding area; it is important to consider patients who are missed due to proximity or other reasons for visiting other hospitals. Second, although we used screening procedures to retain patients living in the main local area as much as possible to ensure that the screening results were closest to the patient’s residence at the monitoring station, individual-level exposure and indoor pollution were still inevitably omitted that could lead to errors in the results. Third, individual characteristics of study subjects such as whether they smoked and drank alcohol, underlying disease states, and household income were not available in the electronic medical records, which prevented us from further analysis of individual differences. Fourth, in addition to the main meteorological factors we studied, other meteorological factors, such as rainfall and duration of UV exposure, may have potential effects on the ocular surface and were not considered. Future multicenter epidemiological studies with larger sample sizes and longer span of periods are still needed.

Conclusion

Overall, we conducted the first time-series analysis of a large sample size covering 59,731 conjunctivitis patients in Urumqi, the farthest from the ocean in the world and the largest city in northwest China (Xinjiang), from January 1, 2013, to December 31, 2020. The association of four major meteorological factors with conjunctivitis outpatient clinics in the region and the lagged effects was revealed by constructing single meteorological factor as well as multimeteorological factor models. We demonstrated for the first time that elevated mean temperature and extremely low relative humidity in Urumqi were risk factors for local conjunctivitis outpatient visits, while elevated atmospheric pressure and extremely low wind speed were protective factors significantly associated with decreased local conjunctivitis risks. In addition, we revealed the U-shaped and inverted U-shaped curves for the dose–response of local temperature and atmospheric pressure on conjunctivitis outpatient visits, with extreme levels (very high or very low) of temperature conditions and extremely low levels of atmospheric pressure also being risk factors for conjunctivitis, providing new evidence in this area of research. Subgroup analysis results suggest that the impact of meteorological factors on conjunctivitis is influenced by gender, age, and seasonal differences. Considering the significant economic burden of conjunctivitis and the impact on the quality of patients’ lives, meteorological factors, especially extreme weather, are strongly associated with the risk of conjunctivitis, which requires larger sample sizes and multicenter epidemiological studies spanning longer periods.