Epidemiological Characteristics of Parainfluenza Virus Type 3 and the Effects of Meteorological Factors in Hospitalized Children With Lower Respiratory Tract Infection

• Published in: Frontiers in pediatrics 2022-04, Vol.10, p.872199
• Main Authors: Xu, Ming, Yue, Wei, Song, Xinyue, Zeng, Luyao, Liu, Li, Zheng, Jinwei, Chen, Xiaofang, Lv, Fangfang, Wen, Shunhang, Zhang, Hailin
• Format: Article
• Language: English
• Subjects: ARIMA; epidemiological characteristics; GAM; HPIV-3; LASSO; meteorological factors; Pediatrics
• ISSN: 2296-2360; 2296-2360
• DOI: 10.3389/fped.2022.872199

To investigate the relationship between meteorological factors and Human parainfluenza virus type 3 (HPIV-3) infection among hospitalized children. All hospitalized children with acute lower respiratory tract infections were tested for viral pathogens and enrolled, at the second affiliated hospital of Wenzhou medical university, between 2008 and 2017. Meteorological data were directly obtained from Wenzhou Meteorology Bureau's nine weather stations and expressed as the mean exposure for each 10-day segment (average daily temperatures, average daily relative humidity, rainfall, rainfall days, and wind speed). The correlation between meteorological factors and the incidence of HPIV-3 was analyzed, with an autoregressive integrated moving average model (ARIMA), generalized additive model (GAM), and least absolute shrinkage and selection operator (LASSO). A total of 89,898 respiratory specimens were tested with rapid antigen tests, and HPIV-3 was detected in 3,619 children. HPIV-3 was detected year-round, but peak activities occurred most frequently from March to August. The GAM and LASSO-based model had revealed that HPIV-3 activity correlated positively with temperature and rainfall day, but negatively with wind speed. The ARIMA (1,0,0)(0,1,1) model well-matched the observed data, with a steady R reaching 0.708 (Ljung-Box Q = 21.178, = 0.172). Our study suggests that temperature, rainfall days, and wind speed have significant impacts on the activity of HPIV-3. GAM, ARIMA, and LASSO-based models can well predict the seasonality of HPIV-3 infection among hospitalized children. Further understanding of its mechanism would help facilitate the monitoring and early warning of HPIV-3 infection.