Association between air pollution and hypertension hospitalizations: a time series analysis in Lanzhou Authors: Anning Zhu, Miaoxin Liu, Jingze Yu, Runping Zhang, Yaoyi Zhang, Rentong Chen, Ye Ruan Published: 2024 ABSTRACT AbstractBackgroundAir pollution is a major environmental risk factor. Hypertension is one of the most important modifiable risk factors for cardiovascular diseases. However, few studies have evaluated the impact of exposure to ambient air pollution on hypertension hospitalizations.ObjectiveThis study aims to explore the correlation between exposure to air pollution and hospital admissions for hypertension, to evaluate the short-term effects of air pollution on hypertension hospitalizations and its lag effects.MethodsWe collected air pollution concentration and hypertension hospitalization data from 2013 to 2020. Distributed lag non-linear models were employed to assess the impact of air pollution on hypertension hospitalizations in Lanzhou City. We also performed subgroup analyses and sensitivity analyses.ResultsA total of 47,884 cases of hypertension hospitalizations were included. Short-term exposure to NO2 and CO increased the risk of hypertension hospitalization. For each 10 µg/m3 increase in NO2 and each 1 mg/m3 increase in CO, the relative risk (RR) for hypertension hospitalization were highest at lag0-3 (RR: 1.0427; 95% CI: 1.0196, 1.0663) and lag3 (RR: 1.0986; 95% CI: 1.0350, 1.1661), respectively. SO2 at different lag structures showed opposite effects on hypertension hospitalization. O3 was significantly negatively associated with hypertension hospitalization. No correlation was observed between PM2.5 and PM10 and hypertension hospitalization. Females and individuals aged ≥ 65 were more susceptible to air pollution. The harmful effects of air pollution are greater in the cold season.ConclusionShort-term exposure to ambient air pollution increases risk of hypertension hospitalization.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12889-024-20740-1.Keywords: Air pollution, Hypertension, Hospitalization, Relative risk Introduction Hypertension is the leading cause of disability and a major risk factor for global mortality. Approximately 10.4 million people die annually due to hypertension [1]. By 2025, it is estimated that the global burden of hypertension will rise by 60%, with the number of hypertensive patients rising to 1.56 billion [2]. In high-income countries, the age-standardized prevalence of hypertension has begun to decline, but in middle- and low-income countries such as China, this proportion continues to rise rapidly [3, 4]. 23.2% of Chinese adults, approximately 244.5 million people, suffer from hypertension [4]. Addressing the burden of hypertension in China is imminent. Identifying risk factors for hypertension can provide crucial information for preventing hypertension and cardiovascular diseases. The onset of hypertension is associated with multiple and complex factors, involving genetics, diet, and various other factors, as well as their interactions. In recent years, the role of air pollution in hypertension has been increasingly emphasized and is considered to be a potential risk factor for hypertension [5, 6]. Air pollution is a major contributor to the global burden of disease, especially in middle- and low-income countries [7]. The Global Burden of Disease Study estimates that air pollution causes approximately 7.1 million premature deaths annually [8]. Air pollution has been found to be strongly associated with hypertension [9]. Each city has its own unique mix of air pollution, meteorological conditions, and social demographics, so findings from studies in other areas may not directly apply to Lanzhou. Furthermore, these studies mainly focus on economically developed regions, with fewer studies conducted in relatively underdeveloped inland areas. Therefore, it is necessary to investigate the correlation between air pollution and hypertension hospitalizations in Lanzhou. Lanzhou City is a transportation hub and industrial center in the northwest region of China. With relatively underdeveloped economy, it is one of the most polluted cities in China. Its unique valley topography makes the dispersion of air pollutants more challenging. The burden of hypertension in Lanzhou is significant, yet there has been no prior research investigating the impact of air pollution on hypertension hospitalizations. This study utilized distributed lag non-linear models (DLNM) to analyze the short-term effects and lag effects of air pollution on hypertension hospitalizations among urban residents. Stratified analyses were conducted based on sex, age, and season. Our aim is to provide scientific evidence for reducing the hospitalization rate of hypertension. Materials and methods Study area Lanzhou City, the capital of Gansu Province, is located in the semi-arid region of northwestern China (latitude 35°34′ to 37°07′ north, longitude 102°36′ to 104°34′ east). As of 2023, the permanent population of Lanzhou is 4.425 million. Lanzhou is a typical river valley basin city, characterized by temperature inversion, dryness, low wind speed, and strong sunlight. Lanzhou is an industrial city primarily focused on petroleum and chemical engineering, machinery manufacturing, metal smelting, and woolen textile industries. Due to its unique geographical and meteorological conditions, the dispersion of air pollutants is restricted in Lanzhou, exacerbating air pollution in the city. Hospitalizations data The hypertension hospitalization data from January 1, 2013, to December 31, 2020, were collected from seven major comprehensive hospitals in Lanzhou City. Based on comprehensive considerations including service quality, medical expertise, and geographical location, this study selected these seven large comprehensive hospitals. These hospitals are widely regarded as the best comprehensive hospitals in Lanzhou City, with over 85% of cardiovascular disease patients choosing to seek treatment there [10]. The collected data includes information such as sex, age, home address, disease codes, admission date, and discharge date. Cases with the ICD-10 code I10 were included in this study after screening using the 10th revision of the International Classification of Diseases (ICD-10). We conducted logical corrections on the inpatient data, removing cases with missing key information such as sex, age, admission, and discharge dates. Based on the registered residential address information of the patients, we screened out patients whose residential addresses were not in Lanzhou, ensuring that the study subjects were permanent residents of Lanzhou City. Air pollution and meteorological data During the study period, air pollution data were obtained from three national air quality monitoring stations in Lanzhou, namely the Gansu Provincial Construction Workers Hospital station, the Railway Design Institute station, and the Biological Products Research Institute station. The locations of the stations are shown in Figure S1. The daily average concentrations of PM2.5, PM10, SO2, NO2, O3, and CO were obtained from the mean values of monitoring results at each station. O3 concentrations are 8-hour daily maximum averages, while concentrations of other air pollutants are 24-hour averages. The national air quality monitoring stations are established according to the National Environmental Protection Standards of the People’s Republic of China (HJ 664–2013), and their locations are far from major roads, industrial pollution sources, and pollutants. Therefore, the monitoring results can effectively represent the overall exposure level of the residents. The daily meteorological data for Lanzhou during the study period were obtained from the China Meteorological Data Service Center (http://data.cma.cn), including daily mean temperature (°C) and relative humidity (%). These two variables were introduced into the model to control for confounding factors. Statistical analysis We described the general characteristics of air pollution concentrations, meteorological factors, and hypertension hospitalizations during the study period. The data were described using measures such as mean, standard deviation (SD), minimum value (Min), maximum value (Max), median, and quartiles. We used Spearman correlation analysis to assess the correlation between air pollutants and meteorological factors. Based on previous studies, the effects of air pollutants on human health exhibit a lag [11, 12]. Additionally, the occurrence of daily hypertension hospitalizations relative to the total population is a rare event, and its distribution approximates a Poisson distribution. Therefore, we employed a combination of Generalized Additive Models (GAM) and Distributed Lag Non-linear Models (DLNM) to investigate the non-linear relationship and lag effects between air pollution concentrations and hypertension hospitalizations. The model is as follows: where t is the observation day; µt refers to the number of hypertension hospitalizations on day t; Log(µt) denotes the expected value of the number of hypertension hospitalizations on day t; α is the intercept; l is the number of lag days; Xt, l denotes the cross-basis matrix obtained by applying DLNM to the six pollutants; β is the matrix coefficient; ns represents natural cubic spline functions; Tempt represents the average temperature on day t; RHt represents the relative humidity on day t; Time represents the time variable; DOW represents the dummy variable controlling for the day of the week effect; Holiday represents the dummy variable controlling for public holidays. This study used the Quasi-Akaike Information Criterion (Q-AIC) to assess model fitting and integrated previous literature to determine the optimal model with the most suitable degrees of freedom (df). A natural spline function with df set to 3 was used to control temperature and relative humidity. The df for the time function was set to 7 to control for long-term trends and seasonality. In our study, we chose the df based on the minimum value of Q-AIC. The smaller value of Q-AIC, the better model fit. Due to the lag effect of air pollution on hypertension, usually within 7 days [13, 14]. In this study, we applied single-day lag (from lag0 to lag7) and cumulative lag (from lag0-1 to lag0-7) to assess the lag effects of pollutants. To investigate the impact of air pollutants on different subgroups of the population, We also conducted stratified analyses based on sex (male and female) and age (< 65 years and ≥ 65 years) [15, 16]. To explore the effects of air pollutants on hypertension during different seasons, we divided the year into warm seasons (April to September) and cold seasons (October to March of the following year) for seasonal analyses [17]. The effect estimates in this study are presented as the relative risk (RR) of hypertension hospitalization per 10 µg/m3 increase in PM2.5, PM10, SO2, NO2, and O3, and per 1 mg/m3 increase in CO, along with their corresponding 95% confidence intervals (95% CI). Finally, to assess the stability of the results, we performed two sensitivity analyses: (1) We simultaneously introduced two air pollutants into the same model to establish a two-pollutant model, in order to test the stability of the model results. (2) We varied the degrees of freedom of the time variable (df = 6–10) and observed the change in RR value. If there were no significant changes, it indicated that the model is robust and the results are reliable. The study was done using R software version 4.2.2 and its “dlnm” package. Statistical tests were two-sided and P < 0.05 was considered statistically significant. Results Description of the general situation Table 1 displays the basic statistical data of air pollutants, meteorological factors, and hypertension hospitalizations. Between January 1, 2013 and December 31, 2020, this study included 47,884 cases of hypertension hospitalizations, with an average of approximately 16 cases per day. 23,708 cases (49.51%) were male, and 24,176 cases (50.49%) were female. 26,415 cases (55.16%) were < 65 years old and 21,469 cases (44.84%) were ≥ 65 years old. The number of hypertension patients admitted was higher in the cold season than in the warm season. The average concentrations of PM2.5, PM10, SO2, NO2, O3, and CO were 49.74 µg/m3, 114.65 µg/m3, 21.64 µg/m3, 45.89 µg/m3, 83.76 µg/m3, and 1.16 mg/m3, respectively. The daily mean temperature and relative humidity during the study period were 11.13 °C and 51.19%, respectively. Table 1Descriptive statistics of air pollutants, meteorological factors and hospitalizations for hypertension in Lanzhou, China, 2013–2020VariableMeanSDMin P 25 Median P 75 Max Air pollutant concentration PM2.5 (µg/m3)a49.7429.270.0031.0042.7760.80278.00PM10 (µg/m3)a114.6583.110.0070.0098.78136.211484.54SO2 (µg/m3)a21.6414.470.0010.8617.5728.60115.51NO2 (µg/m3)a45.8917.950.0034.0044.5654.63146.60O3 (µg/m3)b83.7643.568.0053.0076.00107.00462.00CO (mg/m3)a1.160.690.000.700.961.414.65 Meteorological measures Temp (℃)11.139.88-12.302.2312.6019.7930.40RH (%)51.1915.2911.7140.0051.8562.0096.09 Daily hospitalizations (n) Male8.116.690.002.007.0012.0038.00Female8.276.890.002.007.0012.0043.00Age < 65 years9.047.630.002.008.0014.0043.00Age ≥ 65 years7.356.110.002.006.0011.0038.00Warm16.0912.380.005.0014.0023.0075.00Cold16.6813.490.005.0015.0025.0078.00SD: standard deviation; P25: 25% quartile; P75: 75% quartile; Temp: temperature; RH: relative humiditya 24-hour average; b 8-hour daily maximum averagen represents the number of hypertensive hospitalizations Descriptive statistics of air pollutants, meteorological factors and hospitalizations for hypertension in Lanzhou, China, 2013–2020 SD: standard deviation; P25: 25% quartile; P75: 75% quartile; Temp: temperature; RH: relative humidity a 24-hour average; b 8-hour daily maximum average n represents the number of hypertensive hospitalizations Table S1 displays the correlation coefficients between air pollutants, temperature and relative humidity. PM2.5, PM10, SO2, NO2, and CO are significantly positively correlated with each other. O3 is negatively correlated with PM2.5, PM10, SO2, and CO, with no significant correlation observed with NO2. The temperature is significantly positively correlated with O3 and significantly negatively correlated with other pollutants. Relative humidity shows no correlation with CO and is significantly negatively correlated with other pollutants. The correlation between PM2.5 and PM10 is the strongest (r = 0.850). Figure 1 shows the time series of daily hypertension hospitalizations and air pollutant concentrations from January 1, 2013 to December 31, 2020. During the study period, the number of hypertension hospitalizations showed an increasing trend and fluctuated with the seasons. There were slightly more hypertension hospitalizations during the cold season compared to the warm season. The daily average concentrations of PM2.5, PM10, SO2 and CO generally show a decreasing trend year by year, while NO2 and O3 show an increasing trend year by year. PM2.5, PM10, SO2, NO2, and CO had higher concentrations during the cold season compared to the warm season, while O3 showed the opposite trend. Time series of hypertension admissions and pollutant concentrations in Lanzhou, China, 2013-2020 Short-term effects of air pollutants on hypertension Figure 2 displays contour plots of the RR of hypertension hospitalization across air pollution concentrations and lag days. As pollutant concentrations increase, the risk of hypertension hospitalization increases during specific lag periods, exhibiting a positive correlation. However, in other lag periods, the risk of hypertension hospitalization decreases with rising pollution concentration, exhibiting a negative correlation. The lag effects exhibit different trends depending on the type of air pollutant. Contour plot of relative risk of hypertension hospitalization with changes in air pollutant concentration and lag days Table S2 summarizes the RR (95% CI) for hypertension hospitalizations per 10 µg/m3 increase in PM2.5, PM10, SO2, NO2, and O3 concentrations and per 1 mg/m3 increase in CO concentrations at different lags. NO2 exhibits harmful effects on hypertension hospitalization at lag3, lag0-3, and lag0-4 days, with the greatest effect observed at lag0-3 (RR = 1.0427; 95% CI: 1.0196, 1.0663). SO2 shows harmful effects at lag3, with an RR of 1.0444 (95% CI: 1.0130, 1.0769). In contrast, SO2 is significantly negatively associated with hypertension hospitalization at lag0-2, lag0-6, and lag0-7 days. CO exhibits harmful effects on hypertension hospitalization at lag3 and lag0-3 days, with the largest effect observed at lag3 (RR = 1.0986; 95% CI: 1.0350, 1.1661). O3 is significantly negatively correlated with hypertension hospitalizations at lag0-3, lag0-5 to lag0-7 days. No correlation was found between PM2.5 and PM10 and hypertension hospitalizations. Subgroup analysis Figure 3 and Table S3 presents the RR (95% CI) of hypertension hospitalizations stratified by sex at different lag periods. Among the six pollutants, PM10 and NO2 had a greater effect on hypertension hospitalization in male than in female. PM10 at lag7 (RR = 1.0040; 95% CI: 1.0005, 1.0076), and NO2 at lag0-3 (RR = 1.0465; 95% CI: 1.0216, 1.0721) had the greatest effect on hospitalization for hypertension in male. SO2 and CO had a greater effect on hypertension in female than in male. SO2 at lag3 (RR = 1.0497; 95% CI: 1.0145, 1.0862), and CO at lag3 (RR = 1.1107; 95%CI: 1.0392, 1.1871) had the largest effect on hospitalization for hypertension in female. In both sex groups, there was no correlation between PM2.5 and hypertension hospitalization, while O3 showed a significant negative correlation with hypertension hospitalization. Relative risk (95% CI) of hypertension hospitalization stratified by sex for each 10 µg/m3 increase in air pollutant concentration (each 1 mg/m3 increase in CO) in single-day lags and cumulative lags Figure 4 and Table S4 summarizes the RR (95% CI) of hypertension hospitalization stratified by age at different lag periods. Of the six pollutants, NO2 and CO had a greater impact on individuals < 65 years old. NO2 at lag0-3 (RR = 1.0608; 95% CI: 1.0354, 1.0868), and CO at lag0-3 (RR = 1.1207; 95% CI: 1.0423, 1.2050) had the greatest effect on hospitalization for hypertension at < 65 years of age. SO2 at lag3 (RR = 1.0452; 95% CI: 1.0092, 1.0825) had a greater impact on individuals ≥ 65 years old. In both age groups, there was no correlation between PM2.5 and PM10 and hypertension hospitalization, while O3 showed a significant negative correlation with hypertension hospitalization. Relative risk (95% CI) of hypertension hospitalization stratified by age for each 10 µg/m3 increase in air pollutant concentration (each 1 mg/m3 increase in CO) in single-day lags and cumulative lags Figure 5 and Table S5 displays the RR (95% CI) of hypertension hospitalization stratified by season at different lag periods. SO2, NO2, and CO had a greater impact in the cold season than in the warm season. SO2 at lag3 (RR = 1.0444; 95% CI: 1.0130, 1.0769), NO2 at lag0-3 (RR = 1.0477; 95% CI: 1.0159, 1.0806), and CO at lag3 (RR = 1.1352; 95% CI: 1.0507, 1.2264) had the greatest impact on hypertension hospitalization in the cold season. In both season groups, there was no correlation between PM2.5 and PM10 and hypertension hospitalization, while O3 showed a significant negative correlation with hypertension hospitalization. Relative risk (95% CI) of hypertension hospitalization stratified by season for each 10 µg/m3 increase in air pollutant concentration (each 1 mg/m3 increase in CO) in single-day lags and cumulative lags Sensitivity analysis After selecting the lag day with the maximum RR value for each pollutant in the single-pollutant model, we added additional pollutants to construct the two-pollutant model. The results were compared between the single and two-pollutant models to verify the reliability and stability of the model. Except for CO, the association between pollutants and hypertension hospitalization remained stable and similar to the results of the single-pollutant model (Table S6). Due to the high correlation between PM2.5 and PM10, they were not simultaneously included in the model. We also assessed the robustness of the model by changing the degrees of freedom of the time variable from 6 to 10 (Table S7), and the effect estimates did not change substantially, indicating that the model we constructed is relatively robust. Discussion This study employed DLNM to investigate the correlation between short-term exposure to air pollution and hypertension hospitalization in Lanzhou from 2013 to 2020. The results indicate that NO2 and CO are positively associated with hypertension hospitalization, with significant lag effects. SO2 shows harmful effects in single-day lag structure but is significantly negatively correlated with hypertension hospitalizations in cumulative lag structure. O3 was significantly negatively associated with hypertension hospitalization. In addition, the effect of air pollution on hypertension hospitalization varied by sex, age, and season. Our results further support the notion that air pollution increases the likelihood of hypertension hospitalization. The improvement of air pollution in Lanzhou City from 2013 to 2020 may be the result of efforts to control coal smoke pollution. High levels of particulate matter concentrations occur frequently. The sources of particulate matter are complex, including anthropogenic factors such as fuel combustion, vehicle exhaust, and building construction, as well as natural factors such as dust storms, sand lifting, and floating dust. The increasing trend in NO2 and O3 concentrations may be due to the continued growth in motor vehicle ownership, leading to a large amount of vehicle exhaust emissions. Vehicle exhaust can emit NO2 directly, while O3 is a major secondary pollutant formed through the photochemical oxidation of NO2 and volatile organic compounds emitted by vehicles [18]. The concentrations of PM2.5, PM10, SO2, NO2, and CO are higher in winter and lower in summer. Air pollution is aggravated by the burning of heating fuels in winter, and pollutants are not easily dispersed due to the influence of temperature inversion. O3 exhibits an opposite seasonal pattern, with high concentrations in summer and low concentrations in winter. Ozone concentrations are influenced by solar radiation intensity and temperature, which peaks in summer [19]. In this study, we did not observe a significant statistical correlation between PM2.5 and PM10 and the risk of hypertension hospitalization. This is consistent with some previous research findings. A study conducted in Taipei found no association between PM2.5 levels and the risk of hypertension hospitalization, with odds ratio (OR) = 1.12 (95% CI: 0.95–1.33) on warm days and OR = 1.02 (95% CI: 0.90–1.15) on cool days [20]. A study conducted in Iran found no significant correlation between short-term exposure to PM10 and hypertension hospitalization, with OR = 0.98 (95% CI: 0.95–1.01) [21]. However, some studies have found a positive correlation between short-term exposure to PM and the risk of hypertension hospitalization [22–24]. The inconsistency in PM short-term exposure risk may be due to differences in emission sources, chemical composition, weather conditions, and demographic characteristics, requiring further research to validate this conclusion. This study found a significant positive association between hypertension hospitalization and SO2 exposure at lag3, with RR = 1.0444 (95% CI: 1.0130, 1.0169). At lag0-2, lag0-6, and lag0-7 days, SO2 showed a significant negative correlation with hypertension hospitalization. A study conducted in Taiwan found that short-term exposure to SO2 was associated with a decrease in systolic blood pressure. For every 2.0 ppd change in concentration of SO2 at lag3, there was a change of -0.32 mm Hg in systolic blood pressure (95% CI: -0.62 to 0.03) [25]. However, some studies have also indicated that short-term exposure to SO2 can elevate blood pressure, leading to an increase in emergency visits for hypertension [26–28]. SO2 is a controversial gas. On the one hand, exposure to SO2 may induce systemic inflammation and oxidative damage, which could be one of the potential pathways leading to hypertension [13, 29]. On the other hand, it has been found that SO2 and SO2 derivatives (sulfites and bisulfites) have vasodilatory effects [30, 31]. The vasodilatory effect of SO2 may be related to its mechanism of lowering blood pressure. Further research is needed to elucidate the effect of SO2 on blood pressure. This study showed an adverse effect of NO2 on hypertension hospitalization at lag3, lag0-3, and lag0-4 days, with the greatest effect observed at lag0-3 (RR = 1.0427; 95% CI: 1.0196, 1.0663). Similarly, a study conducted in Xinxiang found that for every 10 µg/m3 increase in NO2 concentration at lag0-4, the RR of hypertension hospitalization was 1.107 (95% CI: 1.046–1.172) [32]. A time-series study conducted in Shijiazhuang found that for every 10 µg/m3 increase in NO2 concentration at lag0-3, there was a 1.18% (95% CI: 0.49–1.87) increase in the hospitalization of hypertensive patients [13]. NO2 is an important air pollutant emitted mainly from vehicle exhaust and stationary sources such as power plants and industrial boilers. Biological mechanism studies have found that the effect of NO2 on blood pressure may be related to oxidative stress, endothelial dysfunction, and inflammatory responses [33]. NO2 exposure upregulates protein biomarkers of vasoconstriction (ET-1, eNOS) and inflammation (TNF-α, IL-1β, ICAM-1), causing vascular endothelial dysfunction [33]. Vascular endothelial dysfunction is closely associated with hypertension [34]. In our study, O3 was significantly negatively associated with hypertension hospitalization at lag0-3, lag0-5 to lag0-7 days. Similarly, a study conducted in China found that each 10 µg/m3 increase in O3 was negatively associated with the risk of hypertension hospitalization at lag0-6 and lag0-7 days, with RR values of 0.986 (95% CI: 0.973, 0.999) and 0.982 (95% CI: 0.968, 0.996), respectively [35]. A study conducted in Canada found that for every 13.42 ppb increase in O3 concentration at lag8, the OR for emergency visits for hypertension in females was 0.869 (95% CI: 0.7444, 0.993) [17]. However, some studies have found a significant positive correlation between short-term exposure to O3 and the risk of hypertension hospitalization [13, 36]. Animal experiments have shown that ozone therapy is an effective method for treating hypertension [37]. Treating hypertensive rats with O2-O3 therapy has been found to reduce systolic pressure, diastolic pressure, and mean arterial pressure. The mechanism involves a decrease in ET-1 levels and an increase in serum renin and NO levels [38]. O3 is a contradictory gas and determining its effect on hypertension remains a major challenge. In this study, CO exhibited harmful effects at lag3 and lag0-3 days, with the greatest effect observed at lag3 (RR = 1.0986; 95% CI: 1.0350, 1.1661). The relationship between CO and hypertension has been widely studied. A cross-sectional study conducted in South Korea found that for every 10 ppb increase in CO, the OR for hypertension hospitalization was 1.171 (95% CI: 1.004, 1.336) [39]. A study conducted in Shijiazhuang found that for every 10 µg/m3 increase in CO at lag0-4, the hospitalization rate for hypertension patients increased by 0.03% (95% CI: 0.01–0.05) [13]. A case-crossover study conducted in Iran found a significant association between every 10 µg/m3 increase in CO and the incidence of hypertension hospitalization, with the best estimate observed at lag6, OR = 1.000076 (95% CI: 1.00002-1.000132) [40]. CO in the environment is mainly produced from incomplete combustion of fossil and biofuels. Mechanisms regarding the deleterious effects of short-term exposure to CO on hypertension have not been fully elucidated, so further experimental studies are necessary. Sex stratification analyses showed that among the six pollutants, PM10 and NO2 had a greater impact on hypertension in males than in females. SO2 and CO had a greater effect on females than males. No harmful effects of PM2.5 and O3 on hypertension hospitalization were observed in both males and females. More and more evidence suggest that males and females may have different health responses to air pollution [41]. Some studies have found a stronger association between air pollution and hypertension in males [6, 42, 43]. However, it has also been shown that air pollution has a greater impact on blood pressure in females [44, 45]. Sex effects may result from a combination of sex-related biological differences and socially derived sex exposures [41]. For example, lung size and growth, particle deposition, gas absorption, blood-gas barrier permeability, airway hyperresponsiveness, vascular reactivity, and inflammation all vary by sex [6]. We suggest that further research could focus on the physiological characteristics of all sexes to explain this difference. Age-stratified analyses showed that of the six pollutants, NO2 and CO had a greater impact on individuals < 65 years old. SO2 had a greater impact on ≥ 65 years. In both age groups, no harmful effects of PM2.5, PM10, and O3 on hypertension hospitalization were observed. Overall, air pollutants had a stronger impact on hypertension hospitalization in those < 65 years old. A study found that short-term exposure to air pollutants had a greater impact on hypertension in young people [6]. This may be related to young people spending more time outdoors, as well as their unhealthy lifestyle habits and emotional states. However, some studies have also indicated that air pollution has a greater impact on cardiovascular diseases in the elderly [10, 46]. It’s possible that the immune systems of older individuals are more susceptible to influence, and their health tends to deteriorate with age, increasing susceptibility to elevated blood pressure or hypertension. Therefore, further research is necessary to investigate the effects of air pollution on different age groups to identify at-risk populations. Seasonal stratification showed that air pollutants (SO2, NO2, and CO) had a higher impact on hypertension hospitalization in the cold season than in the warm season, similar to some studies [13, 26]. This seasonal trend may have the following reasons. Firstly, the emission intensities of SO2, NO2 and CO were significantly higher in the cold season than in the warm season. Lanzhou has a cold winter, winter heating up to five months and mostly use coal as the main energy source of heating. The large combustion of coal results in a significant increase in the concentrations of SO2, NO2, and CO during the winter. Secondly, the unique geographical features of Lanzhou lead to a static wind frequency of 67.4% throughout the year, and the inversion layer exists all year round. Due to strong environmental constraints, industrial and residential heating emissions are not easily dispersed, leading to long-term gather over the city, further exacerbating air pollution [47]. Thirdly, multiple epidemiological studies have shown that cold air stimulation can promote elevated blood pressure [48–50]. Experimental studies have shown that exposure to low temperatures may affect the renin-angiotensin system by upregulating angiotensin receptors, leading to vasoconstriction and elevated blood pressure [47]. Public health interventions can draw inspiration from our results. For example, when outdoor air pollution is severe, especially during cold seasons, educating individuals with hypertension to stay indoors or take protective measures when going outside. This study shows that the lag effect presents different trends with the type of air pollutant. As pollutant concentrations increase, the risk of hypertension hospitalization increases. Even at relatively low concentrations, there may still be an increased risk of hypertension hospitalization. Differences in lag patterns may result from variations in the physicochemical nature of the pollutants, climatic characteristics, and population sensitivity. This result complements the impact of air pollution exposure on hypertension hospitalization among residents in Lanzhou, providing a scientific basis for government departments to issue predictions of air pollution hazards. This study has several strengths. Firstly, we employed a Distributed Lag Non-linear Model (DLNM), combining detailed hypertension hospitalization data and comprehensive air pollution monitoring data, to thoroughly analyze the nonlinear and lag effects of various pollutants on the risk of hypertension hospitalization. This provides new scientific evidence for public health policy formulation. Secondly, to avoid selection bias, our study collected hypertension hospitalization data from seven large comprehensive hospitals in Lanzhou. This allowed us to collect as much case data as possible and therefore may be more representative of the occurrence of hypertension in the general population. However, this study still has some limitations. Firstly, individual exposures were replaced by average ambient air pollution levels from air quality monitoring stations. This alternative method results in a bias in the exposure assessment, which may reduce the accuracy of risk estimation. Secondly, other confounding factors, such as smoking, occupation, outdoor activities, and socioeconomic status, as well as pre-existing diseases and unhealthy factors, were not included, which may lead to bias or other uncertain effects in the results. Conclusion Our study results indicate that short-term exposure to NO2 and CO increases the risk of hypertension hospitalization. SO2 at different lag structures showed opposite effects on hypertension hospitalization. O3 may be associated with a reduced risk of hypertension hospitalizations. No correlation was observed between PM2.5 and PM10 and hypertension hospitalization. Stratified analyses showed that PM10 and NO2 had a greater effect on males, whereas SO2 and CO had a stronger effect on females. People < 65 years old are more susceptible to air pollution. The adverse effects of pollutants are stronger in the cold season. The potential mechanisms of how air pollution affects hypertension need further investigation. Relevant authorities should take active measures to reduce the concentration of ambient air pollution in order to alleviate the disease burden of hypertension. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors acknowledge the dedication and cooperation of all the participants in this study. Author contributions All authors contributed to the study conception and design. AN. Z: Conceptualization, Investigation, Visualization, Writing- Original draft. MX. L: Conceptualization, Writing- Original draft. JZ. Y: Investigation, Writing- Original draft. RP. Z: Software, Writing- Reviewing & Editing. YY. Z: Conceptualization, Visulization. RT. C: Software, Writing- Reviewing & Editing. Y. R: Conceptualization, Writing- Reviewing and Editing, Supervision. All authors reviewed the manuscript. Funding This study was funded by Fundamental Research Funds for the Central Universities, China (lzujbky-2020-9) and Science and Technology Program of Gansu Province, China (20YF3FA027). Data availability The data that support the findings of this study are available from Health Commission of Lanzhou City, a government agency but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Health Commission of Lanzhou City, a government agency. After obtaining third-party permission, data analyzed during the current study are available from the corresponding author (ruany@lzu.edu.cn) on reasonable request. Meteorological data are available at http://data.cma.cn, which is managed by the National Meteorological Information Center of China, but researchers are required to register for free at the site to access the required data. Clinical admission data are not available to the public. R version 4.2.2 software was used in this study, and licenses and other information are available at https://www.R-project.org. The software package used to construct the distributed lag nonlinear model (DLNM) is publicly available at https://www.r-project.org/. Declarations Ethics approval and consent to participate The Institutional Review Board (IRB) at School of Public Health, Lanzhou University has waived the ethics approval and the informed consent for the study. The study adhered to the guidelines and regulations outlined in the Declaration of Helsinki. This exemption is due to the fact that the data used in the study was provided by government agencies and consented for use in this research. The data did not involve any direct or indirect identification of hospital patients’ personal information. And the waiver of informed consent has been reviewed and approved by the research ethics committee. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from Health Commission of Lanzhou City, a government agency but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Health Commission of Lanzhou City, a government agency. After obtaining third-party permission, data analyzed during the current study are available from the corresponding author (ruany@lzu.edu.cn) on reasonable request. Meteorological data are available at http://data.cma.cn, which is managed by the National Meteorological Information Center of China, but researchers are required to register for free at the site to access the required data. Clinical admission data are not available to the public. R version 4.2.2 software was used in this study, and licenses and other information are available at https://www.R-project.org. The software package used to construct the distributed lag nonlinear model (DLNM) is publicly available at https://www.r-project.org/.