Meteorological drivers of second seasonal peak of hand, foot and mouth disease incidence among children and adolescents in Gansu province, 2010 – 2018: a register-based data analysis

  Objective  To assess meteorological drivers of the second seasonal peaks in the incidence of hand, foot, mouth disease (HFMD) among children and adolescents in Gansu province.

  Methods  Totally 83 462 records of 0 – 18 years old HFMD cases registered during 2010 – 2018 were collected from Gansu Provincial Center for Disease Control and Prevention and the data of same period on ambient temperature, relative humidity, rainfall, and wind speed were extracted from the Big Earth Data Platform for Three Poles. Weekly time series were established for the HFMD incidence and meteorological data. Local maximum filter and locally weighted scatter plot smoothing method were used in the identification of bimodal seasonal peaks of HFMD incidence. Logistic regression model analyses were performed to estimate the association of meteorological factors with the occurrence of second seasonal peaks of HFMD incidence, in which, the years with bimodal seasonal incidence peak were considered as the cases and those in years with monomodal incidence peak as the controls and the mean ambient temperature, relative humidity, rainfall and wind speed for 8 weeks before the identified second seasonal peak were considered as the exposure factors.

  Results  Of the 83 190 HFMD cases finally included in the study, 50 663 (60.9%) were male and 77 346 (93.0%) were aged less than 6 years, and more cases were reported in municipality/prefecture of Lanzhou (22 553, 27.1%), Tianshui (9 756, 11.7%), and Dingxi (7 917, 9.5%). Of the 117 observation years for the 13 municipalities/prefecture during the 9 years, 111 (94.9%) of observation years had seasonal peaks of HFMD incidence, of which 92 (78.6%) were with bimodal peak and only 19 (16.2%) were with monomodal peak. The results of logistic regression analysis revealed a strong association between relative humidity of previous one week and the occurrence of second seasonal peak of HFMD, with a 4.2% increase (95% confidence interval 95%CI: 1.5% – 7.3%) in the risk of second seasonal peak occurrence for every 1% increase in weekly average relative humidity; the results also indicated a 193.0% (95%CI: 21.9% – 730.1%) increase in the risk of second seasonal peak occurrence of HFMD incidence for every 1m/s increased average weekly wind during 1 – 7 weeks before the second peak.

  Conclusion  Weekly average relative humidity and wind speed are key meteorological factors correlated with the occurrence of the second seasonal peak of HFMD incidence.