Effects and interaction of temperature and relative humidity on the trend of influenza prevalence: A multi-central study based on 30 provinces in mainland China from 2013 to 2018

Abstract

Background: Evidence is inefficient about how meteorological factors influence the trends of influenza transmission in different regions of China. Methods: We estimated the time-varying reproduction number (Rt) of influenza and explored the impact of temperature and relative humidity on Rt using generalized additive quasi-Poisson regression models combined with the distribution lag non-linear model (DLNM). The effect of temperature and humidity interaction on Rt of influenza was explored. The multiple random-meta analysis was used to evaluate region-specific association. The excess risk (ER) index was defined to investigate the correlation between Rt and each meteorological factor with the modification of seasonal and regional characteristics. Results: Low temperature and low relative humidity contributed to influenza epidemics on the national level, while shapes of merged cumulative effect plots were different across regions. Compared to that of median temperature, the merged RR (95%CI) of low temperature in northern and southern regions were 1.40(1.24,1.45) and 1.20 (1.14,1.27), respectively, while those of high temperature were 1.10(1.03,1.17) and 1.00 (0.95,1.04), respectively. There were negative interactions between temperature and relative humidity on national (SI = 0.59, 95%CI: 0.57–0.61), southern (SI = 0.49, 95%CI: 0.17–0.80), and northern regions (SI = 0.59, 95%CI: 0.56,0.62). In general, with the increase of the change of the two meteorological factors, the ER of Rt also gradually increased. Conclusions: Temperature and relative humidity have an effect on the influenza epidemics in China, and there is an interaction between the two meteorological factors, but the effect of each factor is heterogeneous among regions. Meteorological factors may be considered to predict the trend of influenza epidemic.

Keywords

• Rt
• Influenza
• DLNM
• Meteorological factors
• Multiple random-meta analysis
• Multi-central