Effects of meteorological factors and atmospheric pollutants on the prevalence of respiratory adenovirus in children in Lanzhou, Northwest China

Abstract

Climate change and air pollution have significantly influenced disease prevalence and transmission. Human adenovirus (HAdV) is a common pathogen that causes acute respiratory infections in children. This study employed an ecological approach to investigate the relationship between HAdV infection in children in Lanzhou, Northwest China, and meteorological factors and air pollutants, aiming to deepen the understanding of environmental factors affecting viral transmission. Clinical specimens from children with acute respiratory infections were collected at a sentinel hospital in Lanzhou from January 1, 2023, to February 28, 2025, and were subjected to respiratory adenovirus nucleic acid testing. HAdV infection patterns were analyzed according to age, sex, and season. A combined approach of stepwise linear regression and generalized additive models (GAM) was employed to investigate the correlations between HAdV infection and meteorological factors/air pollutants from January 1, 2023, to December 31, 2024. Between January 1, 2023, and February 28, 2025, a total of 1,339 throat swab specimens were collected from children with acute respiratory infections (ARIs), with a male-to-female ratio of 1.48:1. The HAdV incidence rate in 2024 (11.23%) was higher than that in 2023 (5.45%). The HAdV incidence rates among male and female children were 5.15% and 3.73%, respectively. Preschool-and school-aged children exhibited higher incidence rates than infants and toddlers, at 3.29%, 3.36%, 0.75%, and 1.49%, respectively. HAdV circulated throughout all four seasons, with peak prevalence in autumn (2.28%) and winter (5.06%). HAdV incidence rates showed significant negative correlations with meteorological factors (air temperature, sunshine duration, wind speed), with correlation coefficients of: r = -0.6398 (95% CI: -0.8878, -0.1042), (P = 0.0250); r=-0.6376 (95% CI: -0.8870, -0.1005), (P = 0.0257); r=-0.6208 (95% CI: -0.8809, -0.07285), (P = 0.0312); showed significant positive correlations with atmospheric pollutants (CO, NO₂, and SO₂), with correlation coefficients of: r = 0.7610 (95% CI: 0.3321, 0.9291), (P = 0.0040); r = 0.6846 (95% CI: 0.1824, 0.9035), (P = 0.0140); r = 0.7162 (95% CI: 0.2147, 0.9143), (P = 0.0088); while exhibiting a significant negative correlation with atmospheric pollutant O₃, with a correlation coefficient of r = -0.6938 (95% CI: -0.9067, -0.1992), (P = 0.0123). GAM analysis revealed that meteorological factors (air temperature, sunshine duration, and wind speed) were all nonlinearly associated with adenovirus incidence rates. HAdV was detected at a higher rate in children in the preschool and school-age groups, and most cases were detected in the fall and winter seasons. The incidence rate of HAdV was negatively correlated with meteorological factors (temperature, hours of sunshine, and wind speed) and the atmospheric pollutant O₃, and positively correlated with atmospheric pollutants (CO, NO₂, and SO₂). The influence of these pollutants on the prevalence of HAdV infection should not be ignored.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-24515-5.

Introduction

ARIs are a leading cause of morbidity and mortality in children and infants worldwide. Studies indicate that 80% of ARIs are caused by viruses¹, with HAdV accounting for 2–15% of pediatric respiratory infections^2,3. HAdV belongs to the genus Mammalian Adenovirus of the family Adenoviridae. It is a non-enveloped, linear double-stranded DNA virus with an icosahedral capsid⁴. HAdV can infect both the upper and lower respiratory tracts in children, causing respiratory diseases such as bronchitis, laryngitis, tonsillitis, and pneumonia^5,6. Climate change and air pollution play significant roles in the occurrence and transmission of diseases. By influencing viral epidemiology and population immune status, they can affect the spread and infection of respiratory viruses⁷. This phenomenon is observed globally; however, geographical variations lead to distinct regional climates. Factors such as temperature, daylight hours, wind speed, humidity, and precipitation influence pathogen prevalence^8,9. Multiple studies have demonstrated that air pollutants exert significant adverse effects on respiratory diseases^10–13. In recent years, researchers have identified meteorological factors influencing diseases associated with respiratory infections^14–16. Lanzhou, located in the interior of China, features a dry climate with scarce precipitation and relatively poor natural conditions. This study selected Lanzhou, an inland city in northwestern China, as a representative case. By systematically collecting clinical samples from pediatric acute respiratory infection cases at a sentinel hospital in the Lanzhou region and integrating them with concurrent primary meteorological and air pollutant data, this study aimed to conduct an in-depth analysis of the potential influence of meteorological factors and air pollutants on the epidemiological trends of HAdV infection.

Materials and methods

Overview of the study area

Lanzhou City is located in northwestern China, in the central part of Gansu Province, with its city center at 36°03′ north latitude and 103°40′ east longitude. It is a typical mountainous city in China, with the entire urban area situated in a dumbbell-shaped river valley, surrounded by mountains on all sides. The basin has low wind speeds, with a calm wind rate of 62%. The depth-to-width ratio of the Lanzhou Valley is approximately 0.07, indicating an extremely enclosed mountainous basin terrain. Due to the absence of warm, humid maritime air currents, Lanzhou experiences sparse rainfall throughout the year and has a dry climate. The region receives ample sunlight and has high evaporation rates, characteristic of a typical temperate continental climate.

Specimen source and nucleic acid detection

This study selected pediatric cases of acute respiratory infections from January 2023 to February 2025 at a sentinel hospital in Lanzhou City, Gansu Province, as the study subjects. Inclusion criteria: ① Age ≤ 16 years; ② Symptoms consistent with acute infection (meeting at least one of the following criteria): fever, chills, abnormal white blood cell distribution count (decreased or increased to 5.0 × 10^9/L to 12.0 × 10^9/L); ③ Clinical symptoms (meeting at least one of the following): rhinorrhea, cough with sputum, wheezing, pharyngeal or laryngeal edema or pain, chest tightness or pain, fatigue, abdominal pain, or diarrhea. Qualified medical staff at the sentinel hospital strictly followed the monitoring protocol to collect nasopharyngeal swab samples and gather case data. Specimens must be stored at 4 °C within 24 h of collection and transported to the laboratory. Specimens submitted for testing after 24 h must be stored at -70 °C. This study employed the Xi’an Tianlong Nucleic Acid Rapid Extraction Kit (magnetic bead method) to extract the total viral nucleic acids from the samples. HAdV detection was performed using the HAdV Nucleic Acid Detection Kit developed by Beijing Zhuocheng Huisheng Biotechnology Co., Ltd., in conjunction with the Thermo Fisher Scientific QuantStudio 5 Real-Time Fluorescent Quantitative PCR System. All experimental procedures were strictly followed according to the manufacturer’s instructions.

Meteorological and air quality data collection

Meteorological data are publicly accessible after registration on the official website of the National Meteorological Science Data Center (https://data.cma.cn/). Air quality data were downloaded from the China Air Quality Online Analysis Platform (https://www.aqistudy.cn/). Data includes monthly averages for 2023–2024: Average temperature (°C), Average relative humidity (%) ,Average wind speed (m/s), Average sunshine duration (h), Average precipitation (mm) Monthly averages for temperature, relative humidity, wind speed, sunshine duration, and precipitation were calculated by summing daily meteorological observations for each month and dividing by the number of days in that month. The air pollution indicators included PM2.5, PM10, SO₂, NO₂, CO, and O₃.

Statistical analysis

Data processing and statistical analyses were performed using IBM SPSS Statistics 26.0 software. Normality was assessed using the Shapiro-Wilk test. For continuous variables meeting normality assumptions, between-group comparisons were conducted using independent-sample t-tests. For continuous variables that did not meet the normality assumptions, Mann-Whitney U tests were employed. Between-group comparisons for categorical variables were performed using chi-square or Fisher’s exact tests. Continuous variables with a normal distribution are expressed as mean ± standard deviation (SD). Continuous variables with non-normal distribution are expressed as median (P25, P75). To compare means across multiple groups, univariate analysis of variance (ANOVA) was first performed. If the differences were statistically significant (P < 0.05), Duncan-t was used for pairwise comparisons of the groups. For the correlation analysis, Pearson’s correlation was used if the variables met normality; otherwise, Spearman’s correlation was used. The strength criteria for the correlation coefficients (|r|) were defined as follows: 0 < |r| ≤ 0.2 indicated a weak correlation, 0.2 < |r| ≤ 0.5 indicated a moderate correlation, 0.5 < |r| ≤ 0.8 indicated a strong correlation, and 0.8 < |r| < 1 indicated a very strong correlation. Statistical significance was set at P < 0.05. Figures were generated using GraphPad Prism (version 8.0).

In addition, this study employed a Poisson-based generalized additive model (GAM) to investigate the nonlinear association between meteorological factors and the number of acute respiratory adenovirus (HAdV)-positive cases in children. GAM overcomes the limitations of traditional linear models by utilizing spline functions, which flexibly capture complex nonlinear relationships between the predictor and response variables. Model fitting employed penalized maximum likelihood estimation, with smoothing levels automatically optimized using multiple generalized cross-validation (GCV) from the gcv package. Modeling was based on monthly data spanning multiple months, where the response variable was the number of HAdV infections. The primary predictors included the monthly mean temperature, monthly mean wind speed, and monthly mean sunshine duration. To accurately assess the influence of each factor while capturing the nonlinear relationships between the predictors and response variable, a smoothing term was employed. Cyclic basis splines were used to fit the monthly time series data, thereby correcting for time trends and potential seasonal confounding effects. The GAM model is defined as:

Incidence_rate_i: The observed incidence rate of adenovirus infection in children at time point i (month i). The s() function represents a smoothing spline function, where each smoothing spline employs cubic splines. These reveal a U-shaped relationship: incidence rates are higher at very low and very high temperatures, and lower at moderate temperatures. Threshold effect: The incidence begins to rise rapidly only when temperatures exceed a certain threshold. Saturation effect: Beyond a certain level of increased sunshine duration, the promoting effect on the incidence becomes negligible. Each corresponding degree of freedom (k value) corresponds to a specified maximum value of the wavelet coefficient. The time smoothing terms (time i, k = 4) capture long-term trend changes, thereby reducing the bias caused by time dependence. It controls for confounding: numerous unknown or unmeasured factors (such as population mobility, school openings/closures, co-circulation of other pathogens, and long-term public health policies) vary over time. This term effectively “removes” the overall temporal trend caused by these factors, allowing us to more clearly isolate the “independent” contribution of meteorological factors. It captures seasonality: Respiratory viruses, such as adenoviruses, typically exhibit strong seasonality. While months can partially reflect seasonality, a continuous time smoothing term captures more complex seasonal patterns and non-cyclical, long-term changes.The model exhibited 71.60% interpretable bias, ( AIC = 24.337, R²=0.586, RMSE = 24.651).

Results

Distributional characteristics of children with ARI

From January 1, 2023, to February 28, 2025, 1,339 pediatric acute respiratory cases were collected. Among these, 798 were male children (798/1339, 59.60%) and 541 were female children (541/1339, 40.40%). The male-to-female ratio was 1.48:1. By age group, 194 were infants (≤ 1 year), 288 were toddlers (1–3 years), 422 were preschoolers (3–6 years), and 435 were school-aged children (> 6 years). Seasonal distribution: spring (March–May) 295 cases, summer (June–August) 228 cases, autumn (September–November) 258 cases, and winter (December–February) 485 cases. See Table S1.

Epidemiological characteristics of respiratory HAdV in children

The HAdV incidence rates for January-February 2023, 2024, and 2025 were 5.45%, 11.23%, and 10.33%, respectively, with an overall infection rate of 8.89%. The incidence rate among male children was 5.15%, while that among female children was 3.73%, showing no statistically significant difference between the sexes (P = 0.707). Among the age groups, the preschool group had the highest incidence rate, followed by the school-age group. The rates were as follows: infant group, 0.75%; toddler group, 1.49%; preschool group, 3.29%; and school-age group, 3.36%. The infection rate differed significantly between the > 3 and ≤ 3 years age groups (p = 0.010). Winter was the peak season for HAdV infections. The seasonal incidence rates were as follows: spring, 1.22%; summer, 0.98%; autumn, 2.28%; and winter, 5.06%. Incidence rates in autumn and winter were significantly higher than those in spring and summer (p < 0.01), as shown in Table 1.

Table 1.

Epidemiological characteristics of respiratory adenovirus isolates in Lanzhou City, Northwest China, January 1, 2023 to February 28, 2025.

Characteristics	2023 (n = 514)	2024 (n = 641)	2025* (n = 184)	Total (n = 1339)	p-value (χ²)
Age group, n (%)					0.010 (6.596)
≤ 1 year	1 (0.19)	7 (1.09)	2 (1.09)	10 (0.75)
1–3 years	8 (1.56)	10 (1.56)	2 (1.09)	20 (1.49)
3–6 years	14 (2.72)	27 (4.21)	3 (1.63)	44 (3.29)
> 6 years	5 (0.97)	28 (4.37)	12 (6.52)	45 (3.36)
Gender, n (%)					0.707 (0.141)
Male (cases)	16 (3.11)	42 (6.55)	11 (5.98)	69 (5.15)
Female (cases)	12 (2.33)	30 (4.68)	8 (4.35)	50 (3.73)
Season†, n (%)					< 0.001 (17.679)
Spring	1 (0.18)	14 (1.97)	–	15 (1.22)
Summer	12 (2.15)	0 (0)	–	12 (0.98)
Autumn	9 (1.62)	19 (2.68)	–	28 (2.28)
Winter	29 (5.21)	33 (4.65)	–	62 (5.06)

*Data for 2025 cover only January–February.

^†Seasonal analysis included 1,266 samples (2025 data were included in Winter 2024 according to China’s GB/T 42074 − 2022 seasonal classification standard).

Percentages were calculated as (subgroup cases/total cases in column) × 100%.

P-values were calculated using the chi-square test to compare differences in the classification of each characteristic.

Dashes (–) indicate data not applicable for seasonal analysis of 2025.

Relationship between meteorological factors and HAdV in Lanzhou City

Overview of meteorological elements in Lanzhou City

From 2023 to 2024, the average monthly temperature in Lanzhou City was 11.74 ± 10.37 °C, the average monthly precipitation was 25.70 ± 28.30 mm, the average monthly relative humidity was 49.81 ± 11.38%, the average monthly wind speed was 1.09 ± 0.19 m/s, and the average monthly sunshine hours were 178.07 ± 44.72 h, as shown in Table 2. The temperature in Lanzhou City showed an increasing trend from January to July, with the highest temperature in July, followed by a decline. Relative humidity is low throughout the year, with the highest relative humidity in October, and precipitation peaks in September. The average monthly sunshine hours from May to August are longer than those in other months, and the average monthly wind speed does not change much. The average temperature in Lanzhou varies with the seasons, with the highest in summer and the lowest in winter, reflecting the temperate continental climate of Lanzhou, with cold winters and hot summers, significant temperature differences, low precipitation, and uneven seasonal distribution. The average monthly sunshine hours are significantly longer in spring and summer than in fall and winter, and the average wind speed is higher in spring and summer than in fall and winter. Fig. 1.

Table 2.

Descriptive statistics of meteorological factors and air pollutants in Lanzhou City, Northwest China (January 2023 - December 2024).

Category	Variable	Mean ± SD	Median	Range	P25	P75	Min	Max
Meteorological	Temperature (°C)	11.74 ± 10.37	11.55	30.20	1.60	21.58	− 4.90	25.30
	Humidity (%)	49.81 ± 11.38	48.10	40.60	40.63	57.50	32.00	72.60
	Rainfall (mm)	25.70 ± 28.30	17.40	122.70	8.00	35.53	0.00	122.70
	Sunshine (h)	178.07 ± 44.72	171.00	141.80	147.08	214.58	108.00	249.80
	Wind speed (m/s)	1.09 ± 0.19	1.10	0.70	0.90	1.20	0.80	1.50
Air pollutants	PM₂.₅ (µg/m³)	40.67 ± 14.42	39.50	45.00	27.00	55.00	21.00	66.00
	PM₁₀ (µg/m³)	80.71 ± 41.85	82.00	135.00	53.50	120.25	39.00	174.00
	SO₂ (µg/m³)	13.50 ± 5.33	11.50	21.00	10.00	16.75	7.00	28.00
	NO₂ (µg/m³)	36.46 ± 13.71	31.50	41.00	26.00	51.75	20.00	61.00
	CO (mg/m³)	0.77 ± 0.34	0.62	1.11	0.52	1.03	0.38	1.49
	O₃ (µg/m³)	98.88 ± 35.76	99.50	126.00	70.50	138.25	23.00	149.00

Fig. 1.

Monthly averaged distribution of meteorological elements (Temperature, Relative Humidity, Rainfall, Sunshine Duration, Wind Speed) in Lanzhou, Northwest China, January 2023–December 2024.

Correlation analysis between HAdV and meteorological factors

A study investigated the relationship between HAdV infection and meteorological factors, which was validated using Pearson correlation and regression analyses. Correlation heatmap analysis revealed that HAdV infection was negatively correlated with the monthly average wind speed, monthly average temperature, monthly average sunshine duration, and monthly average precipitation, whereas it was weakly positively correlated with the monthly average relative humidity. Additionally, the monthly average wind speed was positively correlated with the monthly average temperature and sunshine duration, the monthly average temperature was positively correlated with the monthly average precipitation, and the monthly average relative humidity was positively correlated with the monthly average precipitation (S2), indicating complex interactions among these environmental factors. Regression analysis further confirmed these results. The results showed that HAdV incidence rates were significantly negatively correlated with the monthly average temperature, monthly average sunshine duration, and monthly average wind speed, with correlation coefficients of r = -0.6398 (95% CI: -0.8878, -0.1042), (P = 0.0250); r = -0.6376 (95% CI: -0.8870, -0.1005), (P = 0.0257); r = -0.6208 (95% CI: -0.8809, -0.07285), (P = 0.0312) (as shown in Fig. 2a, d, e), the incidence of HAdV showed a positive correlation with the monthly average relative humidity, with a correlation coefficient of r = 0.1793 (Fig. 2b), and a negative correlation with the monthly average precipitation, with a correlation coefficient of r = -0.9986 (Fig. 2c). These findings were highly consistent with the results of the correlation heatmap analysis.

Fig. 2.

Monthly meteorological factors and their correlation with HAdV infection. This figure depicts the association between key meteorological variables and observed HAdV incidence rates from January 2023 to December 2024. Each panel presents the Pearson correlation coefficient (r) for monthly averages of: (a) temperature, (b) relative humidity, (c) precipitation, (d) sunshine duration, and (e) wind speed.

Relationship between atmospheric pollutants and HAdV in Lanzhou City

Overview of atmospheric pollutants in Lanzhou City

The air pollutants in Lanzhou City are PM2.5, PM10, CO, NO₂, SO₂, and O₃. The average concentrations for the two-year period from 2023 to 2024 are as follows: 40.67 ± 14.42 µg/m³; 89.71 ± 41.85 µg/m³; 0.77 ± 0.34 mg/m³; 36.46 ± 13.71 µg/m³; 13.50 ± 5.33 mg/m³; 98.88 ± 35.76 µg/m³, as shown in Table 2. The Air Quality Index (AQI) is an indicator used to measure environmental air quality. The AQI indicates the quantitative severity of primary pollutants, derived from the air quality sub-indices of various air pollutants calculated simultaneously, with the AQI representing the value of primary pollutants. In China, the AQI is divided into six levels: Level 1 is Excellent (0 ~ 50), Level 2 is Good (51 ~ 100), Level 3 is Moderate Pollution (101 ~ 150), Level 4 is Moderate Pollution (151 ~ 200), Level 5 is Severe Pollution (201 ~ 300), and Level 6 is Very Severe Pollution (over 300). The average AQI over the 2-year period was 85.83 ± 15.62. The average concentrations of PM2.5, PM10, CO, NO₂, and SO₂ were higher at the beginning and end of the year and lower in the middle of the year. O₃ showed the opposite trend. PM2.5 had higher average concentrations in January and December; PM10 had higher concentrations from January to April and in December, while other months had relatively lower concentrations; CO had higher average concentrations in January–February and November–December compared to other months; NO₂had higher monthly average concentrations in January–March and October–December, with lower concentrations in July; SO₂ had higher average concentrations in January–March and November–December compared to other months. Unlike other air pollutants, O₃ gradually reached its peak during the middle of the year. The AQI is higher from January to April, with relatively stable values in other months, as shown in Fig. 3.

Fig. 3.

Monthly average distribution maps of environmental factors and air quality index (AQI, including PM2.5, PM10, CO, NO₂, and SO₂) in Lanzhou, Northwest China, from January 2023 to December 2024.

Correlation analysis between HAdV and air pollutants

Correlation heatmap analysis (S3) revealed complex relationships between HAdV infection and the air pollutants. HAdV infection was positively correlated with PM₂.₅, CO, NO₂, and SO₂ concentrations, and exhibited a negative correlation with O₃ concentration. Significant correlations were also observed between pairs of air pollutants. Except for O₃, which showed negative correlations with PM₂.₅, PM₁₀, CO, NO₂, and SO₂. All other pollutants exhibited positive correlations, suggesting potential synergistic effects among different pollutants. Regression analysis further validated these results (Fig. 4). The HAdV incidence rate showed significant positive correlations with CO, NO₂, and SO₂ concentrations, with correlation coefficients of r = 0.7610 (95% CI: 0.3321, 0.9291), (P = 0.0040); r = 0.6846 (95% CI: 0.1824, 0.9035), (P = 0.0140); r = 0.7162 (95% CI: 0.2147, 0.9143), (P = 0.0088) (Fig. 4f, c, d). In contrast, a significant negative correlation was observed with O₃ concentration, with a correlation coefficient of r = -0.6938 (95% CI: -0.9067, -0.1992) (P = 0.0123) ( Fig. 4e). These findings indicate that HAdV infection correlates with elevated levels of certain air pollutants and potentially exhibits a protective association with O₃ concentration.

Fig. 4.

Correlation between monthly air pollutants and HAdV incidence rates. This figure illustrates the associations between key air pollutants and observed HAdV incidence rates from January 2023 to December 2024. Each panel displays the Pearson correlation coefficient (r) for the monthly average of: (a) PM₂.₅, (b) PM₁₀, (c) NO₂, (d) SO₂, (e) O₃, and (f) CO.

GAM analysis results

Based on the collected monitoring data, a GAM was employed to systematically analyze the nonlinear relationship between meteorological factors (monthly average temperature, monthly average sunshine duration, and monthly average wind speed) and the number of respiratory adenovirus infections. The results revealed a significant nonlinear relationship between meteorological factors (monthly average temperature, sunshine duration, and wind speed) and adenovirus infection incidence. The model explained 71.60% of the variance, yielding parameters R² = 0.586 (AIC = 24.337, RMSE = 24.651), indicating the strong predictive capability of meteorological factors (monthly average temperature, monthly average sunshine duration, and monthly average wind speed) for adenovirus transmission risk. The risk peaked at -4.9 °C, where the relative risk was 10.2 times higher than that at the lowest-risk temperature of 25.3 °C ( Fig. 5a). The monthly average sunshine duration also exhibited a significant nonlinear relationship with adenovirus infection rate. The highest infection risk corresponded to 108 h of sunshine, representing a substantial increase compared with the lowest risk at 249.8 h. The peak risk was 1.7 times higher than the minimum value ( Fig. 5b). The monthly average wind speed also exhibited a significant nonlinear association with adenovirus infection rate. Risk peaked at 1.12 m/s, representing a 6.9-fold increase compared to the lowest-risk wind speed of 0.8 m/s ( Fig. 5c). Collectively, these meteorological factors exert a significant influence on adenovirus transmission. Their nonlinear effects suggest that multifactorial integrated assessments are essential for optimizing early disease warning and prevention strategies.

Fig. 5.

GAM risk prediction maps for respiratory adenovirus incidengce rates in Lanzhou, Northwest China, from January 1, 2023, to December 31, 2024, incorporating meteorological factors: (a) monthly mean temperature, (b) monthly mean wind speed, and (c) monthly mean sunshine duration.

Discussion

This study aimed to investigate and analyze the impact of meteorological factors and atmospheric pollutants on the prevalence of HAdV infections among children in Lanzhou, northwest China. The reported HAdV incidence rates for 2023 and 2024 were 5.45% and 11.23%, respectively. While HAdV can affect both the respiratory and gastrointestinal tracts, infections primarily affecting the latter may have been underestimated due to diagnostic errors leading to a lower reported prevalence. These rates are higher than those reported in Hangzhou, China (2.05% and 9.28% for the same years), although both regions showed an increase from 2023 to 2024¹⁷. Potential explanations for this discrepancy include sampling error, or alternatively, this period (2023–2024) represents the post-COVID-19 era. In 2023, following the relaxation of strict control measures and concurrent with the pandemics of SARS-CoV-2 and influenza¹⁸, strong public awareness of prevention, including continued mask-wearing and hand hygiene (non-pharmaceutical interventions, NPIs), may have suppressed HAdV transmission. The subsequent increase in HAdV positivity in 2024 could be attributed to a relaxation of individual NPI compliance, potentially increasing HAdV transmission risk. In this study, HAdV incidence rates did not show a statistically significant difference between genders (P = 0.707). Among age groups, infants and toddlers exhibited significantly lower infection rates compared to preschoolers and school-aged children (P = 0.010), a trend consistent with global reports^19–21. This observation might be associated with the gradual lifting of NPIs and increased inter-personal contact among school-aged children upon returning to school, facilitating HAdV spread. With regard to seasonality, incidence rates were significantly higher in autumn and winter compared to spring and summer (P < 0.01), a finding consistent with the research by Qianwen Zhao et al. in Guangdong Province²².

Lanzhou, the capital of Gansu Province, is located in the semi-arid region of northwestern China (longitude 102°35′55″ to 104°34′29″E, latitude 35°34′20″ to 37°07′N) in a typical river valley basin, with an arid climate with little rainfall, strong evaporation, low wind speeds, a high frequency of static winds, and a deep inversion layer, which is characterized by a temperate semi-arid climate²³. As an important petrochemical, metallurgical, and mechanical industrial base, Lanzhou has long suffered from severe air pollution. Although effective control measures have been implemented since 2013, the PM2.5 concentration is still approximately 1.14 times the national annual standard limit (35㎍/m³).

Several theories have proposed an association between meteorological factors and respiratory viruses^24,25. Meteorological factors influence respiratory viruses through four primary mechanisms: viral survival, infection, transmission, and host immune responses. First, meteorological factors can alter viral stability, viability, activity, pathogenicity, or virulence, thereby affecting their survival time. Laboratory studies indicate that most viruses can survive longer at lower temperatures and relative humidity levels²⁶. Some aerosol viability experiments suggest that adenoviruses are more stable at high relative humidity levels²⁷. This study employed a combination of stepwise linear regression analysis and GAMs to reveal correlations between meteorological factors, atmospheric pollutant indicators, and HAdV incidence rates. HAdV exhibited a significant negative correlation with the monthly average temperature, with higher infection rates observed at colder temperatures. This may be attributed to the decline in temperature during colder weather, which reduces the body’s physical barriers, decreases respiratory mucosal secretions, and weakens immune function, thereby increasing the risk of infection. The risk of HAdV infection peaked when the monthly average temperature was − 4.9 °C, suggesting the need for enhanced preventive measures. Consequently, children are more susceptible to HAdV infections in Lanzhou during winter months with lower temperatures, consistent with the findings of Tang et al. ²⁸. HAdV showed a significant negative correlation with the monthly average sunshine duration. GAM predictions revealed that longer sunshine durations (249.8 h) were associated with lower HAdV incidence rates, coinciding with the lowest risk of HAdV infection incidence. This could be due to sufficient sunlight promoting higher evaporation rates, leading to a dry climate that affects the survival of HAdV. GAM predictions also identified a peak in HAdV risk under specific meteorological conditions when the monthly average sunshine duration was approximately 108 h, highlighting the need for intensified monitoring and preventive measures during such periods. HAdV showed a significant negative correlation with the monthly average wind speed. GAM predictions indicated that the relative risk of childhood respiratory adenovirus infection peaked at wind speeds of 1.12 m/s, signifying an increased transmission risk and necessitating particular attention and control measures. This may be related to the influence of calm conditions on HAdV transmission, whereas strong winds might affect viral deposition.

Ambient air pollution is considered a significant risk factor for viral respiratory infections, with complex mechanisms likely involving the induction of inflammatory responses, promotion of cell death, induction of oxidative stress, and regulation of viral receptor expression, among others²⁹. Our findings revealed a strong association between elevated air pollutant levels and HAdV incidence rates. Ambient air pollutants are broadly categorized into gaseous pollutants and total suspended particulates (TSPs), with gaseous pollutants including CO, SO₂, NO₂, and O₃, and TSPs primarily comprising PM2.5 and PM10. Studies have indicated that exposure to air pollutants can trigger the generation of reactive oxygen species (ROS), leading to oxidative stress, increased mucus secretion, and cytokine production, thereby adversely affecting lung health^30,31. Furthermore, air pollution impairs host defense mechanisms. Exposure to air pollutants reduces the phagocytic capacity of macrophages and compromises immune function³². Collectively, these factors enhance the susceptibility of the population to respiratory pathogen infections. Previous research has shown that SO₂ is a known inducer of respiratory inflammation, causing direct airway damage and disruption of respiratory barrier function^33,34. Our study results indicated a positive correlation between CO, NO₂, and SO₂ and the incidence rate of HAdV. Conversely, a significant negative correlation was observed with O₃, suggesting that high concentrations of CO, NO₂, and SO₂ increase HAdV infection, whereas high concentrations of O₃ decrease it. This finding contrasts with a study in Shenyang³⁵, which reported an increase in respiratory mortality associated with rising O₃ levels, suggesting a potential paradox. Such discrepancies may reflect the specificity of pollutant effects on different respiratory diseases or pathogens. Second, regional variations may have contributed to this phenomenon. Our study area differs from previously studied regions in terms of air pollution composition, demographic structure, and prior health status, factors that may lead to different characteristics in the correlation between pollutants and respiratory diseases. Finally, it is important to note that research on the relationship between air pollutants and respiratory viral infections is still in its nascent stage and exhibits considerable heterogeneity. Future studies should incorporate data from multiple locations and various virus types to further delineate these relationships and underlying mechanisms. The positive correlation between SO₂ and HAdV incidence rates is consistent with a study conducted in Wuhan, China³⁶. Therefore, the influence of meteorological factors and atmospheric pollutants on the incidence rate of pediatric HAdV should not be disregarded. During periods of extreme weather events or high air pollution, more proactive and appropriate measures should be implemented to protect children from developing respiratory infections.

This study had several limitations. First, the research period was relatively short, and it focused solely on HAdV as the pathogen. Consequently, our exploration of the correlations between meteorological factors, air pollutants, and HAdV infection rates was preliminary. Our findings are primarily based on statistical associations; the intrinsic mechanisms by which meteorological factors and air pollutants influence HAdV infection may be more complex and require further research. Second, the association between meteorological factors, air pollutants, and HAdV may be influenced by various potential confounding factors that were not fully controlled or incorporated into this study’s analysis. This may limit our ability to obtain more comprehensive findings. Furthermore, this study was conducted in a single city, which may limit the external validity (generalizability) of its findings, as it does not account for broader regional variations. Given these limitations, future research should prioritize extending the study duration, incorporating a wider range of respiratory pathogens, and expanding the geographical scope of the investigations. This will contribute to uncovering more granular and universally applicable findings, thereby providing a more robust foundation for a deeper understanding of the role of environmental factors in respiratory viral infection.

Conclusion

In summary, this study thoroughly investigated the potential impact of meteorological factors (temperature, sunshine duration, and wind speed) and atmospheric pollutants (CO, O₃, NO₂, and SO₂) on the prevalence of HAdV infection in children in the Lanzhou region. The results showed that the incidence rate of HAdV was relatively high among preschool and school-age children, with a significant concentration during autumn and winter, revealing a close association between environmental factors and HAdV incidence. The GAM model established in this study revealed a quantitative relationship between key meteorological factors and HAdV incidence. This relationship can be utilized to construct a meteorological data-based HAdV incidence risk-prediction model. When temperatures persistently fall below specific thresholds, sunshine duration decreases, and specific wind speed conditions are present, the system can automatically trigger intermediate- or high-level alerts. Public health authorities can integrate this warning system into existing respiratory infectious disease surveillance networks, enabling weeks of advanced risk prediction. This provides a scientific basis for medical resource allocation and public alerts. Following high-risk signals from the warning system, tiered and targeted NPIs measures can be activated, avoiding blanket lockdowns to minimize societal costs of the latter.This study emphasizes the need to strengthen targeted public health intervention measures during periods of low temperatures and high pollution to effectively protect children’s health. This study not only fills the gap in research on HAdV epidemiology and its environmental risk factors in the Lanzhou region but also provides valuable data to support the development of more precise and effective public health strategies under different climate and pollution conditions in the future. By elucidating how meteorological factors and changes in air pollutants may influence the occurrence of HAdV infections in children, our research provides important scientific evidence and practical guidance for optimizing disease prevention and control measures and formulating relevant health policies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

B. W. method design, experimental manipulation, first draft writing, software processing, review and editorial writing; H. Z. experimental manipulation, software processing, data management, first draft writing; X. Z. program design, obtaining grants; M. D. program design, project management; S. L. program design, project management; H. W. experimental manipulation; M. W. experimental manipulation.All authors read and approved the final manuscript.

Funding

This study was supported by the Gansu Provincial Key Research and Development Program—Social Development Field Special Project ( No.: 23YFFA0051) and the Gansu Provincial Disease Prevention and Control Research Project (No.: GSJKKY2025-44).

Data availability

All data geneerated or analyzed during this study are included in this published article.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

This study was approved by the Ethical Committee of Gansu Provincial Center for Disease Control and Prevention, and carried out strictly in accordance with the Declaration of Helsinki. Informed consent was obtained from all participants and from their legal guardians who were aged ≤ 16 years. Data were stored and analyzed anonymously. Inclusion criteria: ①age ≤ 16 years old; ②symptoms consistent with acute infection (at least one of the following): fever, chills, abnormal white blood cell distribution count (decreased or increased);③clinical symptoms (at least one of the following): runny nose, coughing and sputum, wheezing, pharyngeal and laryngeal edema or soreness, chest tightness and chest pain, fatigue, abdominal pain and diarrhea. Nasopharyngeal swab specimens were collected by qualified medical staff of the sentinel hospitals in strict accordance with the monitoring program, and case information was collected. Specimens were stored at 4 ℃ for 24 h after collection and transported to the laboratory, and specimens sent for examination for more than 24 h were stored at -70℃.

Consent to participate

Informed consent was obtained from the parents or guardians of all participants. The parents or guardians were informed of the laboratory results of pathogen detection.