Wuxi, located in East China in the southern region of Jiangsu Province, lies within the Yangtze River Delta. It borders the Yangtze River to the north and Taihu Lake to the south. Characterized by a northern subtropical humid monsoon climate, Wuxi covers a total area of 4627.46 km². As of the end of 2019, Wuxi city had a permanent resident population of 6.5915 million.

The daily varicella case data for Wuxi from January 1, 2010, to December 31, 2019, were sourced from the Wuxi Center for Disease Control and Prevention (CDC). Wuxi city follows a comprehensive infectious disease reporting protocol in compliance with the National Infectious Disease Surveillance System (NDSS) requirements. Medical facilities are required by law to report varicella cases online within 24 hours, providing details such as gender, age, diagnosis date, and hospital visitation information. Stringent data collection, reporting, and inspection procedures ensure the quality of the data. Since June 1, 2017, varicella has been classified as a class C infectious disease under the management of Jiangsu Province. Varicella case data are collected and counted based on the name of the disease, rather than its classification. Therefore, the change in the classification of varicella as an infectious disease before and after June 1, 2017, does not affect the statistical results.

Meteorological data for Wuxi city from January 1, 2010, to December 31, 2019, including temperature (°C), average surface temperature (°C), relative humidity (%), atmospheric pressure (hPa), wind speed (m/s), rainfall (mm), evaporation (mm), and sunshine duration (h), were sourced from the Wuxi City Meteorological Bureau.

The data were organized using MS Excel 2016 (Microsoft Office), and all data analysis was performed using R 3.6.2 (The R Foundation for Statistical Computing) with a significance level of α=.05. Quantitative variables were described statistically using the mean (SD) and quantiles (0%, 25%, 50%, 75%, and 100%). Count data were described statistically in terms of frequency and composition ratio, that is, n (%). A time series graph illustrating varicella case numbers and fluctuations in meteorological factors was generated. Additionally, Spearman correlation analysis was conducted to assess the relationships between variables.

The Generalized Additive Model (GAM) is a widely used approach for investigating the influence of meteorological factors on health outcomes. For instance, in a multicenter study, the GAM was used to estimate the risk of pulmonary tuberculosis associated with meteorological factors in eastern China. The study revealed that low temperatures, relatively high wind speeds, and low relative humidity contribute to an increased risk of pulmonary tuberculosis [18]. In Hong Kong, GAM analysis revealed that both rotavirus and norovirus hospitalizations were closely related to recent precipitation changes, although in opposite directions [19]. To our knowledge, this is the first study to apply the Poisson GAM and the Piecewise Linear Regression Model to explore the relationship between 7 climate variables and the incidence of varicella in Wuxi, in East China’s Huadong Region.

The GAM incorporates the following covariates: First, a smoothing term is generated using a thin plate spline function with a maximum degree of freedom of 6 to account for long-term trends in the disease number sequence. Second, the number of cases from the previous day to address short-term autocorrelation in the incidence sequence. Third, a binary indicator for holidays, including weekends and statutory holidays. Fourth, seasonal variables to manage the seasonal fluctuations in disease incidence. Subsequently, we compared the generalized cross-validation scores of models across varying lag durations and identified the lag durations with the minimum score as the optimal one for subsequent analysis [20]. Following this, exposure-response curves were plotted to illustrate the relationship between meteorological factors and varicella incidence risk at the optimal lag durations. A segmented regression function was used to assess the magnitude of change in the effect of unit changes in meteorological factors. Additionally, subgroup analysis was conducted to evaluate the robustness of the association. The effect size was quantified by the percentage change in varicella incidence risk, along with its corresponding 95% CI, for every 1-unit increase in meteorological factors.

This study was approved by the Institutional Review Board of Wuxi Hospital Affiliated with Nanjing University of Chinese Medicine (2024 (paper)-031-01). The study utilizes varicella monitoring data from the Wuxi Center for Disease Control and Prevention, which have been anonymized to ensure effective protection of patient privacy.

Table 1 presents the summary statistics of meteorological factors and varicella cases in Wuxi city from January 1, 2010, to December 31, 2019. A total of 64,086 cases of varicella were reported during the study period. Males accounted for 54.16% (n=34,709 cases), while females accounted for 45.84% (n=29,377 cases). Teenagers, children, and infants under the age of 17 years comprised the largest group, representing 78.37% (n=50,221 cases) of the total reported cases.

Figure 1 illustrates the time distribution of daily varicella cases and meteorological variables from 2010 to 2019. According to the monthly chart of varicella incidence between 2010 and 2019 (Figure 2), the highest incidence of varicella occurs from November to January of the following year, accounting for 42.54% (27,261/64,086) of all cases. By contrast, May to June shows a smaller peak, with reported cases making up 18.11% (11,605/64,086) of the total cases.

According to the Spearman correlation coefficient (Table 2) and the Spearman matrix pie chart (Figure 3) of meteorological factors and varicella cases in Wuxi from 2010 to 2019, the Spearman correlation coefficient between air pressure and varicella cases is 0.145. This indicates a positive correlation, suggesting that the number of varicella cases tends to increase with higher atmospheric pressure. Conversely, temperature, average surface temperature, sunshine duration, evaporation, and wind speed exhibit negative correlations with varicella cases, with coefficients of −0.130, −0.124, −0.058, −0.169, and −0.202, respectively. This suggests that as these meteorological factors increase, the number of varicella cases tends to decrease.

We compared the generalized cross-validation scores of models with different lag days and selected the lag days corresponding to the minimum score as the optimal lag days (Table 3). We then plotted the exposure-response curve between meteorological factors and the risk of varicella incidence for the optimal lag days (Figure 4) and used a segmented regression function to evaluate the magnitude of effect changes for unit changes in meteorological factors. Subgroup analysis was conducted to assess the robustness of the association (Table 4).

From 2011 to 2018, Wuxi city had the highest incidence of varicella in Jiangsu Province [21]. Since December 2018, the city has included the varicella vaccine in its routine immunization program, providing 2 free doses to eligible resident children [22]. Despite a reduction in varicella incidence in 2019, with a reported rate of 174.04 per 100,000 population, Wuxi city’s rate remained significantly higher than the national average of 70.14 per 100,000 population.

Previous studies have demonstrated that climatic factors such as temperature, humidity, rainfall, and wind speed significantly influence the occurrence of certain infectious diseases [23]. Our study has identified a nonlinear relationship between the risk of varicella (chickenpox) incidence and various climatic factors, including temperature, air pressure, precipitation, and relative humidity. Specifically, as temperature, surface mean temperature, and air pressure increase, the risk of varicella incidence initially rises and then declines. By contrast, with increasing precipitation and relative humidity, the risk of varicella initially decreases and then increases. Additionally, evaporation and wind speed are negatively correlated with varicella risk, while sunshine duration shows a positive correlation. In Wuxi city, varicella cases predominantly involve males, children, and students, aligning with trends observed in other regions of China [9]. However, in tropical regions, adolescents and young adults are more susceptible to varicella virus infection [11]. Additionally, our study identified a bimodal seasonal pattern of varicella incidence, consistent with patterns observed in Jinan [15], Shandong, and Lu’an [24], Anhui, China. This contrasts with studies conducted in Japan [17] and the United Kingdom [25], where the bimodal periodicity of varicella incidence in northern Japan transitions to a unimodal pattern in southern Japan, especially at lower latitudes. In the United Kingdom, varicella incidence also follows a bimodal pattern, with peaks observed in mid-March and June. These variations underscore the potential influence of location-specific climatic factors on the transmission of the VZV. Such discrepancies suggest that regional climatic conditions play a significant role in shaping the patterns of varicella incidence across different geographical areas.

Our research confirms previous findings regarding the relationship between temperature and varicella incidence risk. We observed that as both temperature and surface mean temperature increase, the risk of varicella incidence initially rises and then declines. This pattern is consistent with the results of a study conducted in Guangzhou, China, which also identified a ∩-shaped curve relationship between average temperature and varicella risk [8]. Our study found that when the average temperature and surface mean temperature are ≤20°C and 22.4°C, respectively, the risk of varicella incidence increases as these temperatures decrease, especially among male children under 17 years. The high incidence of varicella during the winter and spring seasons can be attributed to the significant temperature fluctuations during these periods. Temperature variations can weaken the immune system, increasing the risk of varicella infection. Specifically, the cold and dry weather in winter, combined with gradually rising temperatures and large temperature fluctuations in spring, creates favorable conditions for the transmission and reproduction of the VZV. During cold winter days, people tend to gather indoors more frequently, which increases interpersonal contact. As varicella is primarily transmitted through respiratory droplets and direct contact, such gatherings facilitate the spread of the virus. The study by Markus et al [26] further corroborates these findings, indicating that the VZV exhibits heightened replication efficiency at lower temperatures.

The exposure-response curve between air pressure and the risk of varicella exhibited a nonlinear relationship. Initially, as air pressure increased, the risk of varicella rose before eventually declining. Specifically, when the air pressure was ≤1011.28 hPa, a decrease in air pressure corresponded with an increased risk of varicella. These findings align with results from a study conducted in Guangzhou, China, which also demonstrated a similar pattern in the relationship between air pressure and varicella risk [8]. Although a clear causal relationship between changes in air pressure and the incidence of varicella has not been widely confirmed, variations in air pressure may indirectly influence the incidence of varicella by affecting human physiological functions and the immune system.

Our study found that the “exposure-response” curves for precipitation and relative humidity in relation to varicella incidence risk were consistent, both showing a nonlinear pattern. As precipitation and air humidity increased, the risk of varicella incidence initially decreased and then increased, particularly among children under 17 years. A study in Jinan also reported a positive correlation between varicella incidence and rainfall [15]. Another study showed that rainfall can indirectly influence varicella incidence by altering relative humidity and human activities [27]. As a nonlipid-coated virus, the VZV is more likely to survive and spread under specific humidity conditions. Relevant studies have shown that the risk of varicella peaks at 69.5% relative humidity [28]. This suggests that the varicella virus can maintain its activity and increase the risk of infection in high-humidity environments. Therefore, during rainy weather or when the varicella virus is particularly active, it is important to enhance cleanliness and air circulation in the environment. Methods such as opening windows for ventilation, using ultraviolet radiation, and other disinfection techniques can help reduce the risk of varicella virus transmission.

Our study found a positive correlation between sunshine duration and the risk of varicella, which aligns with findings from a study conducted in Jinan. Previous research has indicated that differences in varicella incidence between tropical and temperate regions are primarily influenced by sunshine hours and ultraviolet radiation [29]. A study conducted in Poland found that solar ultraviolet radiation during the summer might inhibit cellular immunity, leading to an increased incidence of herpes zoster [30]. Additionally, solar radiation may impact vitamin D metabolism and the host’s immune system, thereby elevating the relative risk of varicella [31]. Biological evidence indicates that vitamin D can inhibit human natural killer cells, which are crucial for innate defense against viral infections [32].

The associations between wind speed, evaporation, and varicella incidence were nearly linear. As wind speed and evaporation increased, the risk of varicella incidence gradually decreased. Wind affects the suspension time and diffusion distance of the virus. Sabariego et al [33] demonstrated that higher wind speeds can dilute local virus concentrations due to the resuspension phenomenon.

Additionally, global climate change, extreme weather events, population movements, changes in air pollutant concentrations, and vaccination rates may impact varicella incidence [34]. A systematic review of Latin America and the Caribbean found that the incidence of varicella has been increasing due to urbanization and higher population density [35].

This study has several limitations that should be noted. First, as an ecological study, it may be subject to ecological fallacies. The meteorological data used represent the overall averages for Wuxi city, and cannot be matched precisely to the specific meteorological conditions experienced by each case. As a result, accurate individual exposure levels could not be determined. This means that the research results may not be directly applicable to individual-level inferences. Second, due to the limited monitoring data available, the model only considered meteorological factors. Other potential influences on varicella incidence, such as host susceptibility, population density, air pollutants, and socioeconomic factors, were not included in the analysis. These limitations suggest that while the research offers valuable insights into the relationship between meteorological factors and varicella, further studies are needed to address these limitations and provide stronger data to support varicella prevention and control strategies. Additionally, this study is confined to environments similar to subtropical climates or urban areas and may not be generalizable to other regions.

In this study, we identified significant correlations between varicella incidence and specific climatic factors in Wuxi from 2010 to 2019. Notably, varicella cases increased during months with high temperatures and humidity. By monitoring and analyzing meteorological data, we can predict trends and peak periods of varicella incidence, allowing for timely issuance of early warning information to alert relevant departments and the public to take preventive measures. This approach can help reduce varicella incidence and mitigate the impact of outbreaks on socioeconomic activities and daily life. While our study offers valuable insights, we recognize that the sample size and data collection period may limit the generalizability of our findings. Future research should explore broader geographical areas and longer timeframes to further validate our results and investigate other climatic factors that may affect varicella transmission.