Shenzhen Influenza Analysis:Time-space Distribution,Excess Mortality,Meteorology And Ambient PM_2.5 Concentration Implications

Objective:1.To discover the characteristic and spatial and temporal distribution of influenza in Shenzhen during 2013-2015 through analyzing influenza monitoring data2.Provide scientific reference for disease control and prevention through discovering the influence of meteorological factors and ambient PM_2.5 concentration on the incidence of influenza in Shenzhen City.3.Provide scientific reference for understanding the epidemic of influenza and formulating the strategy of vaccination through using Serfling regression to estimate the excess mortality caused by influenza in Shenzhen during 2013-2015.Methods:1.Shenzhen Influenza data during 2013-2015 period were collected from 18 surveillance sites.Spatial-temporal scan statistic were utilized to analyze the temporal and spatial distribution characteristics of influenza.2.Shenzhen influenza-like illness（ILI）record,meteorological data and PM_2.5 data during 2013-2014 period were collected.Scatter plot and generalized additive model are used to analyze the relationship between meteorological factors,ambient PM_2.5 concentration and incidence of influenza.3.Age-specific Serfling regression model,which is established based on the data of all cause deaths（AC）,respiratory and circulatory deaths（R&C）and pneumonia and influenza deaths（P&I）,is used to estimated the excess mortality and death numbers attributed to influenza for three age groups in Shenzhen: the all-age,< 65 years old and ≥65years old.Results:1.The number of ILI reported were 378685 during 2010-2015.The total number of outpatient visits was 10085716,and the percentage of influenza-like-illness（ILI%）was about 3.75%.Each year,there are 2 ILI% peaks,March to July and December to February.People suffered from ILI were mainly from 0-4 age group.With age increased,the constituent ratio decreased gradually.2.Temporal-spatial scanning statistics showed that the high morbidity area of ILI was mainly from April to June.The low morbidity area of ILI are mostly in September and October.The high morbidity aggregation time was consistent with the descriptive analysis of ILI.Further analysis found that some areas kept high morbidity for over 8 month,which indicates long-term trend of ILI.3.The number of throat swab samples collected were 8854 in Shenzhen during 2013-2015,and 1064 influenza viruses were identified by virus isolation and identification.The positive rate was 12.02%.433 isolates（40.70%）were of influenza A subtype H3N2.339 isolates（31.86%）were of influenza A（H1N1）pdm09.291 isolates（27.35%）were of influenza B virus.1 isolates（0.09%）were of influenza A subtype H1N1.The positive rate of virus had statistical difference among different year（X2=63.4424,P<0.0001）.The influenza positive rate in 2015 is lower that those in 2013 and 2014.As for the most popular dominant strain of influenza in Shenzhen,they were influenza A（H1N1）pdm09 and influenza A subtype H3N2 in 2013 and 2014,and influenza A subtype H3N2,influenza B（Yamagata）virus in 2015.The dominant strain showed a mixed epidemic of two strains in recent three years.And there is an inverse trend in different subtype,among which influenza A（H1N1）pdm09,influenza A subtype H3N2 andinfluenza B（Yamagata）virus showed an active performance.4.The influenza peak in Shenzhen mainly occurred when fullfing weather condition such as weekly lowest temperature at 10℃ to 25℃,relative humidity at 70% to 85%,weekly rainfall at 0 to 150 mm,weekly average temperature difference at around 5.5℃ and weely average PM_2.5 10 to 30μg/m3.Generalized additive model display that 1℃ increment of minimum temperature per week was associated with the influenza-like illness values of（RR）0.9922（0.9908-0.9936）for all age group,1% increment of relative humidity was associated with the influenza-like illness values of（RR）1.0019（1.0011-1.0026）,1mm increment of total precipitation per week was associated with the influenza-like illness values of（RR）1.00048（1.00040-1.00056）for all age group,1℃ increment of mean temperature difference was associated with the influenza-like illness values of（RR）1.0107（1.0056-1.0157）,An IQR（29.3 ug/m3）increment of ambient PM_2.5 concentration was associated with the influenza-like illness values of（RR）1.0291（1.0180-1.0403）for all age group.5.According to AC-based estimates,the average excess death number due to influenza in Shenzhen from year 2013-2014 to year 2015-2016 is 652.8,and the yearly excess death rate is 6.11/100000.Base on R&C,it is estimated that the average excess death number due to influenza is 449.78,and the yearly excess death rate is 4.21/100000.Finally,the P&I method indicates that the average excess death number due to influenza is 44.94,and the yearly excess death rate is 0.42/100000.Conclusion:1.Shenzhen influenza monitoring data during 2013 to 2015 period indicates active influenza performance mainly in summer and winter,and those who suffer from influenza most are age between 0-14 years old.In the meantime,the spatial distribution of influenza is different in Shenzhen from 2013 to 2015,and the influenza has circulated for a longer time in some areas.2.There may be a certain correlation between meteorology factor（temperature,relative humidity,rainfall,temperature difference）,ambient PM_2.5 concentration and influenza in Shenzhen.Meteorology factor and ambient PM_2.5 concentration are influenced on ILI in Shenzhen.3.The Serfling model shows that the elderly,especially those already suffer from respiratory and circulatory system diseases are the high-risk groups of excess death due to influenza.