Temporal and spatial characteristics of highly pathogenic avian influenza outbreaks in China during 2004 to 2015

Temporal and spatial characteristics of highly pathogenic avian influenza outbreaks in China during 2004 to 2015 Mingyue Liu,∗ Qian Lu,∗,1 Shuxia Zhang,† Xiaolong Feng,∗ and Md. Shakhawat Hossain∗ ∗

College of Economics and Management, Northwest Agriculture and Forestry University, Yangling 712100, Shaanxi, China; and † College of Veterinary Medicine, Northwest Agriculture and Forestry University, Yangling 712100, Shaanxi, China outbreaks had obvious seasonal effects clustered in January to February, June, and November. Spatial characteristics revealed that outbreaks were more frequent in Xinjiang, Hubei, and Guangdong but caused a larger number of dead animals in Liaoning and Shanxi. HPAI H5N1 appeared in 25 provinces, while HPAI H5N2 was mainly localized in Hebei and Jiangsu, and HPAI H5N6 occurred in Heilongjiang, Jiangsu, Hunan, and Guangdong. HPAI viruses were most frequently detected in chickens and wild birds in northern China, while the majority of HPAI infections were identified in chickens, ducks, and geese in southern China. Regionally, HPAI outbreaks were most frequent in the western region but resulted in larger number of animals dying or being culled in the eastern region. These findings could provide a new understanding of the distributional characteristics of HPAI outbreaks and offer prospects for better prevention and control strategies.

ABSTRACT Identifying the temporal and spatial characteristics of highly pathogenic avian influenza (HPAI) outbreaks is very important for developing effective and appropriate countermeasures against HPAI and promoting sustainable development in the poultry industry. This study aimed to analyze four aspects of the temporal and spatial characteristics of HPAI outbreaks in China, including the frequency of HPAI outbreaks, numbers of dead animals (died or culled), types of HPAI viruses, and species of infected animals. Temporal characteristics showed that the frequency of HPAI outbreaks decreased and then increased, with some years deviating from the main trend in 2004 to 2010 and 2011 to 2015, while the largest number of dead animals due to HPAI outbreaks was in 2005. During 2004 to 2015, HPAI H5N1 was the major type of HPAI virus, and chickens had the greatest risk of being infected with HPAI, followed by ducks and geese. The HPAI

Key words: temporal characteristic, spatial characteristic, highly pathogenic avian influenza, China 2017 Poultry Science 0:1–9 http://dx.doi.org/10.3382/ps/pex097

INTRODUCTION

addition, human infections with HPAI H5N1, H7N9, H5N6, and H10N8 have been reported in mainland China (Gao et al., 2013), and the public has reduced consumption of poultry products due to fear of HPAI. To control and prevent HPAI outbreaks, China has taken several important steps, including measures such as stamping out, movement control, cleaning and disinfecting infected premises, and adoption of a nationwide massive vaccination campaign combined with intensive post-vaccination surveillance efforts (Martin et al., 2011). Nevertheless, timely and effective prevention and control measures should be based on explicitly analyzing and summarizing the distributional characteristics of HPAI to more effectively deal with future outbreaks. Therefore, identifying the temporal and spatial characteristics has important practical significance for improving prevention and control measures targeted to reduce the risk of HPAI outbreaks. Temporal and spatial characteristics could provide clues that assist in understanding the dynamics of disease spread. Detection of spatial, temporal, and spacetime clustering have been found to be essential in

China is one of the primary poultry producers in the world. According to statistical data published by the Food and Agriculture Organization (FAO) in 2013, Chinese poultry production generates 18.94 and 29.13 million tons of meat and eggs, respectively (FAO, 2013). In China, poultry production has assumed an important position in the domestic and international arena and played a vital role in developing the agricultural economy and increasing the income of rural people. However, China is among the countries most affected by highly pathogenic avian influenza (HPAI) (Wang et al., 2013). Since the first major emergence in Guangxi in 2004, outbreaks associated with HPAI H5N1 have been reported throughout China, with more than 100 reported outbreaks and more than 40 million poultry slaughtered (Martin et al., 2011; Li et al., 2015). In  C 2017 Poultry Science Association Inc. Received October 10, 2016. 1 Corresponding author: [email protected]

1

2

LIU ET AL.

identifying higher risk areas where, and times when, disease surveillance and control should be targeted (Si et al., 2008). A number of studies have been conducted to analyze the spatial-temporal patterns of global and regional HPAI H5N1 outbreaks. From November 2003 to March 2007, a total of 3,345 H5N1 outbreaks were reported worldwide; these outbreaks had significant spatial clustering and were mainly concentrated in South Asia, East Russia, West Europe, the Black Sea, and East Africa and also showed clear seasonal patterns, with a high density of outbreaks in winter and early spring (Si et al., 2008, 2009). Zhang et al. (2014) investigated the global distribution and spread of HPAI H5N1 outbreaks from March 2003 to March 2012 using geographic information system (GIS) methods. They identified a total of 5,963 HPAI H5N1 outbreaks; most outbreaks were reported in Asia (61.16%), Europe (21.95%), and Africa (16.89%), and the countries with the highest frequencies and rates on these continents were Vietnam, Germany and Romania, and Egypt, respectively. The distribution of outbreaks demonstrated obvious clustering patterns during the periods from March 2003 to March 2006 and August 2009 to March 2012; however, clustering patterns occurring from April 2006 to July 2009 were not as obvious. Several previous studies have analyzed the distributional patterns of HPAI H5N1 in regions such as Vietnam, Bangladesh, and Indonesia; these studies have found that the occurrence, spread, and distributional patterns of disease varied within different spatial contexts, with the epidemic showing a bipolar distribution in Vietnam, radial distribution in Bangladesh, and a shrinking pattern in Indonesia (Lee et al., 2015; Zhang et al., 2015). In China, the distribution of HPAI H5N1 outbreaks from 2004 to 2006 showed a clustered pattern (Oyana et al., 2006); however, outbreaks took place sporadically after the 2006 poultry vaccination campaigns (Li et al., 2015), did not occur in 2010, and then increased again after 2010 (Huang and Wang, 2015). The outbreaks revealed significant seasonal patterns and mostly occurred during the cold season (Li et al., 2015). The periods during which HPAI H5N1 outbreaks took place were different in different animals: wild bird outbreaks tended to concentrate in January, February, April, and May; poultry outbreaks concentrated in January, February, June, and November; and human cases were identified predominately in January (Tian et al., 2015). Research studies on this topic have always used spatial analysis methods to detect distribution patterns of global or regional HPAI H5N1 outbreaks. These studies provided valuable observations, evidence, and information that formed the basis of our study. However, existing research has concentrated only on the HPAI H5N1 virus and ignored the HPAI H5N2 and H5N6 outbreaks that have occurred in recent years. Additionally, most previous studies were focused on changes in the frequencies of HPAI outbreaks, and only a minority of studies have considered the number of dead animals and types of viruses and species of infected animals

involved. Finally, findings from earlier studies did not contain recent observational data and could not reflect the latest changes in HPAI outbreaks. In this study, we attempted to analyze the frequency of outbreaks, numbers of dead animals, types of viruses, and species of infected animals involved in outbreaks during 2004 to 2015 to characterize explicitly the temporal and spatial characteristics of HPAI outbreaks and provide essential information for developing appropriate and effective countermeasures against HPAI.

MATERIALS AND METHODS Data The first major outbreaks of HPAI H5N1 in mainland China started in January, 2004, and as the outbreaks unfolded, the disease was detected widely throughout the country (Martin et al., 2011). Hence, data from HPAI outbreaks reported since 2004, including the numbers of HPAI outbreaks and dead animals, types of HPAI viruses, and species of infected animals, were used to analyze the temporal and spatial patterns of HPAI outbreaks in China. This study was based on secondary data collected mainly from the Official Veterinary Bulletin published on the website of the Chinese Ministry of Agriculture and official emergency management reports issued by the same ministry during 2004 to 2015. Missing data were obtained from the internet and news reports.

Methods Frequency of HPAI Outbreaks and Numbers of Dead Animals The frequency of HPAI outbreaks and numbers of dead animals were two important indicators that reflected distributional outbreak patterns. The number of outbreaks during a period was used as an indicator of the frequency. The specific formulas used are as follows. The interannual and intermonthly frequencies of HPAI outbreaks in China are:

BCt =

12 31  

Xtij

(1)

j =1 i =1

BCi =

31 2015  

Xtij

(2)

j =1 t =2004

Where BCt and BCi are the interannual and intermonthly frequencies of HPAI outbreaks, respectively, t = 2004, 2005, · · · , 2015, and i = 1, 2, · · · , 12, j = 1, 2, · · · , 31. Xtij is the number of HPAI outbreaks  in ith month of t year in j province; 12 i =1 Xtij is the  HPAI outbreaks in t year in j province; 2015 t =2004 Xtij is the HPAI outbreaks in ith month in j province.

CHARACTERISTICS OF HPAI OUTBREAKS

The interannual and intermonthly numbers of dead animals due to HPAI outbreaks in China are:

BQt =

31  12 

Ytij

(3)

j =1 i =1

BQi =

31 2015  

Ytij

(4)

j =1 t =2004

Where BQt and BQi are the interannual and intermonthly numbers of dead animals, respectively, t = 2004, 2005, · · · , 2015, and i = 1, 2, · · · , 12, j = 1, 2, · · · , 31; Ytij is the number of dead animals in ith month  of t year due to HPAI outbreaks in j province; 12 total number of dead animals i =1 Ytij is the  in t year in j province; 2015 t =2004 Ytij is the total number of dead animals in ith month in j province. Types of HPAI Viruses and Species of Infected Animals Influenza viruses in the family Orthomyxoviridae are classified into three types: A, B, and C. Based on antigenic differences in viral surface glycoproteins, type A influenza viruses are further classified into different hemagglutinin (HA) and neuraminidase (NA) subtypes (Capua and Alexande, 2009; Jiang et al., 2012). Currently, 16 HA subtypes (H1-H16) and 9 NA subtypes (N1-N9) of influenza virus have been isolated from birds (Tong et al., 2012), and among these subtypes, H5 and H7 have been found to be highly virulent in poultry. The types of HPAI outbreaks in mainland China during 2004 to 2015 included H5N1, H5N2, and H5N6, of which HPAI H5N1 was the major type, as H5N1 outbreaks occurred more frequently and caused greater damage to poultry farms than other subtypes (Huang and Wang, 2015). Avian influenza viruses have been found to exist in a wide variety of poultry (turkey, chicken, guinea fowl, quail, goose, duck, etc.) and wild bird species (geese, ducks, swans, tern, etc.) throughout the world but have mainly been identified in migratory waterfowl, especially ducks. However, the costs associated with avian influenza virus infection in turkeys and chickens have been the highest (Suarez, 2000). Temporal and Spatial Characteristics Analysis Temporal characteristics were explored by estimating the variations trend in frequency and numbers of dead animals, types of HPAI viruses, and species of infected animals in HPAI outbreaks. In previous studies, simple linear regression models have been used to analyze variations in trends, including those of extreme temperature, rainfall, meteorological drought, and consecutive days of severe weather. However, HPAI outbreaks had a nonlinear tendency with a declining tendency at first and an increasing tendency thereafter (Huang and Wang, 2015). Thus, a simple linear regression model was not appropriate for the analysis of HPAI outbreaks, and a nonlinear regression equation was adopted to analyze the trends of HPAI outbreaks in this study. The

3

model is expressed by the following equation:

y ( t ) = b 0 + b1 t + b2 t 2

(5)

Where y (t) is the frequency of HPAI outbreaks or numbers of dead animals, types of HPAI virus, or species of infected animals, t is the time of HPAI occurrence during 2004 to 2015 and b0 , b1 , and b2 are parameters to be estimated. Thus, in the equation y  (t) = b1 + b2 t, y  (t) reflects the trend in HPAI outbreaks at a given time. When y  (t) > 0, y (t) has an increasing trend. When y  (t) < 0, y (t) has a decreasing trend, and y  (t) = 0 indicates a turning point. Additionally, spatial characteristics were determined by analyzing the nationwide and regional distribution of HPAI outbreaks. The spatial distribution of HPAI outbreaks was analyzed in ArcGIS-ArcMap version 9.0. supplied by Environmental Systems Research Institute (ESRI)

RESULTS AND ANALYSIS Temporal Characteristics of HPAI Outbreaks Interannual Characteristics of HPAI Outbreaks During 2004 to 2015, 133 HPAI outbreaks were reported in China, including 123 H5N1 outbreaks, 3 H5N2 outbreaks, and 7 H5N6 outbreaks. These outbreaks affected many animals, such as chickens, ducks, geese, wild birds, and quail, resulting in a total of 32.03 million animals that died or were culled. The interannual characteristics of HPAI outbreak frequency are displayed in Figure 1A. A total of 50 HPAI outbreaks occurred in 2004, and outbreaks decreased during 2004 to 2009, there were no outbreaks in 2010 and sporadic outbreaks since. In addition, our study used nonlinear regression models to identify trends in the frequency of outbreaks. The R2 value for the binary regression model for interannual characteristics was 0.9114 and significant at the 1% level; therefore, the frequency of HPAI outbreaks had a fluctuant and significant trend that decreased during 2004 to 2010 and increased during 2011 to 2015. The number of dead animals was 2.95 million in 2004, jumped rapidly to 22.27 million in 2005, and then declined sharply to 3.00 million in 2006. The figure showed a trend of ups and downs in the subsequent years, the lowest figure was zero in 2010 and the relative high figure was 1.61 million in 2012 (Figure 1B). Overall, the numbers of dead animals showed no obvious trend and was not consistent with the trend in outbreak frequency; this lack of trend was possibly because of differing densities of animals in different periods and provinces. We also used a nonlinear regression model to identify trends in the number of dead animals; however, the R2 value for this model was 0.3235 and therefore not significant. Hence, the number of dead animals had no obvious tendency.

4

LIU ET AL.

Figure 1. Interannual characteristics of HPAI outbreaks in China during 2004 to 2015. The frequency of HPAI outbreaks, numbers of dead animals, types of HPAI viruses, and species of infected animals are presented in 1A, 1B, 1C, and 1D, respectively. The columns show annual values and the dotted lines are the trend lines.

During 2004 to 2012, only HPAI H5N1 outbreaks occurred in China, and no HPAI H5N2 or HPAI H5N6 outbreaks were reported (Figure 1C). HPAI H5N1 outbreaks occurred during each year from 2004 to 2015, except for 2010; this strain accounted for 92.48% of the total HPAI outbreaks. HPAI H5N2 outbreaks first occurred in 2012 and then occurred 2 times in 2005, accounting for 2.26% of the total outbreaks. HPAI H5N6 occurred 1 and 6 times in 2014 and 2015, respectively; HPAI H5N6 outbreaks accounted for 5.26% of the total outbreaks. As shown in Figure 1C, we identified an obvious trend in HPAI H5N1 outbreaks; however, the trends in HPAI H5N2 and H5N6 outbreaks were not significant. The result of the binary regression model for H5N1 supported the aforementioned analysis; the R2 value for this model was 0.8898 and significant at the 1% significance level, indicating that HPAI H5N1 outbreaks had decreasing and increasing trends in 2004 to 2010 and 2011 to 2015, respectively. Since 2004, HPAI viruses have been detected in chickens, ducks, geese, wild birds, quail, and other birds. This meant that the aforementioned mentioned poultry were susceptible to HPAI virus infections (Huang and Wang, 2015). In 2004 to 2009, 2011, and 2015, HPAI infections were identified in at least two types of animals; however, during 2012 to 2014, the HPAI viruses were detected only in chickens (Figure 1D). Of the infected species, HPAI viruses were most commonly

detected in chickens; a total of 105 chicken infections were identified, accounting for 73.43% of the total outbreaks, and the proportion of total outbreaks in duck and goose infections reached 13.29% and 7%, respectively. HPAI viruses were identified in wild birds only in 2006, 2009, and 2015 and in quail and rare birds in 2006 and 2015, respectively. Only chicken HPAI infections during 2004 to 2015 had an obvious trend, and the R2 for this model was 0.8491 and significant at the 1% significance level. This meant that the number of chicken identified with HPAI virus infections decreased and increased with fluctuation in 2004 to 2010 and 2011 to 2015, respectively. Intermonthly Characteristics of HPAI Outbreaks Outbreaks of HPAI were mainly concentrated in February and January during 2004 to 2015; outbreaks in these months occurred 40 and 25 times, respectively, and together accounted for 48.88% of the total outbreaks (Table 1). This trend is possibly because the weather was cold and frequent trade of poultry occurred during the Spring Festival, increasing the risk of disease transmission. There were 23, 9, and 7 outbreaks of HPAI concentrated in November, June, and April, respectively; 6 outbreaks were identified in December, and 5 outbreaks were detected in May. Li et al. (2015) similarly found that outbreaks of HPAI in domestic poultry mainly took place in 4 months during the cold season (November, December, January, and February), while

5

CHARACTERISTICS OF HPAI OUTBREAKS Table 1. Intermonthly characteristics of HPAI outbreaks in China during 2004 to 2015. Month

Frequency (times, %)

Numbers of dead animals (million, %)

Province

1

25 (18.80%)

0.87 (2.72%)

15

H5N1 (22), H5N2 (1), H5N6 (2)

2 3 4 5 6 7 8 9 10 11 12

40 4 7 5 9 2 4 5 3 23 6

3.15 (9.83%) 0.46 (1.45%) 0.40 (1.26%) 0.08 (0.25%) 3.21 (10.01%) 0.52 (1.62%) 0.35 (1.10%) 0.61 (1.90%) 0.14 (0.44%) 21.88 (68.31%) 0.35 (1.11%)

16 3 5 4 8 2 4 4 3 9 6

H5N1 H5N1 H5N1 H5N1 H5N1 H5N1 H5N1 H5N1 H5N1 H5N1 H5N1

(30.08%) (3.01%) (5.26%) (3.76%) (6.77%) (1.50%) (3.01%) (3.76%) (2.26%) (17.29%) (4.50%)

Types of HPAI

(39), H5N6 (1) (4) (6), H5N6 (1) (5) (9) (2) (2), H5N2 (1), H5N6 (1) (4), H5N6 (1) (3) (23) (4), H5N2 (1), H5N6 (1)

only a few appeared in warm months (July, August, and September). The numbers of dead animals due to HPAI outbreaks in November, June, February, and January, were 21.88, 3.21, 3.15, and 0.87 million, respectively. Greater numbers of provinces were affected by HPAI outbreaks in February, January, November, and June, with 16, 15, 9, and 8 provinces affected during these months, respectively (Table 1). Hence, a greater risk was associated with HPAI in these 4 months (January, February, June, and November), when more provinces experienced HPAI outbreaks and outbreaks were more frequent and had a larger number of dead animals. HPAI H5N1, H5N2, and H5N6 outbreaks occurred in January, August, and December. HPAI H5N1 and H5N6 outbreaks were identified in January, April, and September, and HPAI H5N1 outbreaks took place in the rest of the months. Overall, H5N1 was the major type of HPAI virus identified, occurring in all 12 months; however, most H5N1 outbreaks took place in February. The next most frequently identified virus outbreaks were H5N6 and H5N2, observed in 6 and 3 months, respectively (Table 1). HPAI viruses were detected in a single animal in March and July and identified in at least two animals during the other months. Of the infected animals, chickens, ducks, and geese were most susceptible to HPAI virus infection, and infections in these species were identified during eight different months. Wild birds were identified to have HPAI viruses only in April, May, January, and February; this trend may be associated with the migration of birds during the spring and autumn. HPAI infections were detected in quail and rare birds in January, June, and December (Table 1).

Spatial Characteristics of HPAI Outbreaks Nationwide Distribution of HPAI Outbreaks During 2004 to 2015, HPAI outbreaks occurred in 26 provinces in China, resulting in approximately 32.03 million destroyed animals; only the Beijing, Shandong, Fujian, Chongqing, and Hainan provinces were not af-

Types of infected animal Chickens (16), ducks (6), geese (3), wild birds (1), quail (1) Chickens (39), wild birds (1) Chickens (4) Chickens (3), geese (1), wild birds (3) Chickens (1), ducks (1), geese (1), wild birds (2) Chickens (6) ducks (2), geese (1), rare birds (1) Chickens (2) Chickens (2), ducks (1), geese (1) Chicken (4), ducks (1) Chickens (3), ducks (2), geese (1) Chickens (21), ducks (3), geese (1) Chickens (4), ducks (2), geese (1), rare birds (1)

fected. The number of dead animals due to HPAI outbreaks in Liaoning was the largest, with approximately 20.07 million; in other provinces, the number of dead animals was less than 1.56 million. HPAI H5N1 was the major type of HPAI virus, and chickens had the highest risk of HPAI virus infection. Among these 26 provinces, the number of HPAI outbreaks in Xinjiang, Hubei, and Guangdong was highest, with 13 to 16 outbreaks in each province, and in Hunan and Tibet, each with 9 to 12 outbreaks per province. In Anhui, Yunnan, Liaoning, Ningxia, Jiangsu, Jiangxi, and Guizhou, 5 to 8 HPAI outbreaks were identified, and in the remaining 14 provinces, the number of outbreaks was less, at approximately 1 to 4 outbreaks per province (Figure 2A). Overall, HPAI outbreaks were more frequent south of the Yangtze River than in the north of China. The majority of HPAI outbreaks in southern China were concentrated in Guangdong, Hubei, and Hunan, whereas in northern China, outbreaks mainly occurred in Xinjiang, Ningxia, and Liaoning. The numbers of dead animals due to HPAI outbreaks were larger in Liaoning and Shanxi, with 1.51 to 20.07 million in these provinces, followed by Ningxia, Xinjiang, Guangdong, and Hunan, with 1.01 to 1.50 million, and Yunnan, Hubei, Gansu, Guizhou, Anhui, and Hebei, with 0.51 to 1.00 million; however, in the remaining 14 provinces, the numbers were lower, with approximately 0.01 to 0.50 million animals dying or being culled (Figure 2B). Therefore, the number of dead animals caused by HPAI viruses in northern China was higher than the number in southern China. Of the types of HPAI viruses, HPAI H5N1 outbreaks occurred in 25 provinces, but not in Heilongjiang, during 2004 to 2015 and occurred most frequently in Xinjiang (Figure 2C). HPAI H5N2 was most frequently identified in Hebei and Jiangsu, with 1 and 2 outbreaks each, respectively. HPAI H5N6 outbreaks were mainly concentrated in 4 provinces (Heilongjiang, Jiangsu, Hunan, and Guangdong); of these provinces, the number of outbreaks in Hunan was the highest. Overall, Jiangsu was the only province where HPAI H5N1, H5N2, and H5N6 occurred, and it is one of the most

6

LIU ET AL.

Figure 2. Nationwide distribution of HPAI outbreaks in China during 2004 to 2005. The frequency of HPAI outbreaks, numbers of dead animals, types of HPAI viruses, and species of infected animals are presented in 2A, 2B, 2C, and 2D, respectively.

7

CHARACTERISTICS OF HPAI OUTBREAKS

vital provinces in terms of poultry production. Therefore, the local government should strengthen surveillance for HPAI viruses to prevent and control the spread of HPAI outbreaks. HPAI viruses were identified in chickens in 23 provinces but not in Qinghai, Zhejiang, and Shanghai; among these 23 provinces, Xinjiang experienced the most outbreaks (Figure 2D). HPAI virus infections in ducks were mainly concentrated in Inner Mongolia, Xinjiang, Shanghai, Jiangxi, Hunan, Hubei, Sichuan, Tibet, Guangdong, and Guangxi. Most of these provinces are in the south of China, and this spatial distribution might be associated with the water bodies and feeding habits in these provinces. HPAI viruses were identified in geese in Xinjiang, Jiangsu, Zhejiang, Anhui, Hubei, Tibet, and Guangdong, most of which are southern provinces. Wild birds with HPAI were identified in Qinghai, Ningxia, Liaoning, Henan, and Tibet, possibly because these provinces occur along migratory bird flyways, including East Africa-West Asia, Central Asia, and East Asia-Australian flyways, along which wild birds can spread HPAI viruses through contaminated water, contact with poultry, and so forth (Fang et al., 2008). HPAI virus infections in quail were most frequently identified in Guizhou, and infections in rare birds were most frequently identified in Jiangsu and Hunan; however, infections of these species were rarely identified. Regional Distribution of HPAI Outbreaks Poultry production is mainly concentrated in the eastern region of China, followed by the central and western regions. However, the number of HPAI outbreaks in the western region of China was highest, followed by the central and eastern regions (Table 2). This meant that HPAI outbreaks might be associated with poultry feeding patterns that vary in different regions. One possible explanation was that free-range poultry production systems and poor conditions for preventing and controlling diseases existed in the western region, while poultry were usually bred in industrialized farms with good husbandry practices and were vaccinated in the central and eastern regions of mainland China. The numbers of dead animals resulting from HPAI outbreaks were dissimilar in different regions. The largest number of dead animals was identified in the

eastern region of China, with 22.16 million; this pattern might be associated with the higher poultry density and more frequent trade of live poultry in the eastern regions, which had a smaller number of more impactful outbreaks. In the western and central regions of China, HPAI outbreaks were associated with 5.61 and 4.26 million dead animals, respectively (Table 2). The HPAI H5N1 virus appeared in the eastern, central and western regions of China, while HPAI H5N2 was most frequently identified in the eastern region of China, and HPAI H5N6 occurred in the eastern and central regions of China (Table 2). This pattern meant that the eastern region of China had a greater variety of HPAI virus outbreaks, whereas the central region of China had two types of HPAI virus outbreak, and the western region of China had outbreaks of only one type of virus. Hence, local governments in the eastern region of China should take more initiative to collect animals and conduct virological surveillance. Within the three regions of China, species of infected animals only differed slightly. Chickens, ducks, geese, and wild birds were all identified to have HPAI viruses in these regions. Specifically, chicken and wild bird, duck, and goose HPAI infections were identified most frequently in the western, central, and eastern regions, respectively. Rare birds were most frequently identified to have HPAI virus infections in the central and eastern regions, and infected quail were mainly identified in the western region of China (Table 2).

DISCUSSION Since the first outbreak was officially reported in Guangxi Province in 2004, HPAI has been detected widely throughout China. Infections caused by HPAI have caused heavy losses and posed significant threats to humans and poultry. In this study, we analyzed four aspects of the temporal and spatial characteristics of HPAI outbreaks in China, including the frequency of outbreaks and numbers of dead animals, types of HPAI viruses, and species of infected animals, to provide some suggestions to the government regarding effective HPAI outbreak prevention and control. As HPAI outbreaks caused severe damage to the poultry industry, some studies have been conducted to

Table 2. Regional distribution of HPAI outbreaks in China during 2004 to 2015. Region

Frequency (times, %)

Numbers of dead animals (million, %)

Types of HPAI

Eastern region1

31 (23.30%)

22.16 (69.19%)

H5N1 (25), H5N2 (3), H5N6 (3)

Central region2

46 (34.59%)

4.26 (13.31%)

H5N1 (42), H5N6 (4)

Western region3

56 (42.11%)

5.61 (17.50%)

H5N1 (56)

Types of infected animal Chickens (20), wild birds (2), Chickens (36), wild birds (1), Chickens (49), wild birds (4),

ducks (3), geese (5), rare birds (1) ducks (10), geese (2), rare birds (1) ducks (5), geese (3), quails (1)

1 The eastern region of mainland China includes Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan. 2 The central region of mainland China includes Shanxi, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, and Hunan. 3 The western region of mainland China includes Sichuan, Chongqing, Guizhou, Yunnan, Tibet, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang, Guangxi, and Inner Mongolia.

8

LIU ET AL.

identify the temporal and spatial distributions of global HPAI outbreaks (Si et al., 2008, 2009; Li et al., 2014; Zhang et al., 2014), and other studies focused on certain regions or a single country, such as Asia (Gilbert et al., 2008), Africa (Cattoli et al., 2009), Thailand and Vietnam (Pfeiffer et al., 2007; Tiensin et al., 2007), Bangladesh and Indonesia (Loth et al., 2010; Zhang et al., 2014), and Romania (Farnsworth and Ward, 2009). Some studies have analyzed the distribution of HPAI outbreaks in China (Oyana et al., 2006; Martin et al., 2011; Huang and Wang, 2015; Li et al., 2015; Tian et al., 2015); however, these studies all focused on the distribution of HPAI H5N1 outbreaks, and no studies have attempted to identify the temporal and spatial characteristics of the HPAI H5N2 and H5N6 outbreaks that have caused damage in recent years. In addition, data previously used to analyze the distribution of HPAI outbreaks did not contain recent data and, therefore, could not reflect the distribution of outbreaks in recent years. The numbers of outbreaks have increased in recent years, and 12 outbreaks were reported in 2015, which indicates that these outbreaks should be taken into account when analyzing the temporal and spatial characteristics of HPAI outbreaks. Finally, most previous studies assessed the number of HPAI outbreaks, and only a few researchers have considered numbers of dead animals and types of HPAI viruses and species of infected animals. In our study, data from 2004 to 2015 were used to analyze the temporal and spatial characteristics of HPAI outbreaks in China. Temporal characteristics were identified through determination of the interannual and intermonthly characteristics of HPAI outbreaks. The interannual characteristics showed that the frequency of HPAI outbreaks decreased during 2004 to 2010 and then increased in 2011 to 2015. The number of dead animals due to HPAI outbreaks was higher during 2004 to 2006, and the highest number of dead animals was identified in 2005, with 22.27 million. After 2006, the number reduced to zero in 2010 and increased in later years. In 2004 to 2012, only HPAI H5N1 occurred in China, and after 2013, HPAI H5N2 and H5N6 began to appear and then increased with fluctuation. The species of infected animals included chickens, ducks, geese, wild birds, quail, and rare birds; of these species, chickens had the highest risk of HPAI virus infection, followed by ducks and geese. These intermonthly characteristics demonstrated that HPAI outbreaks in domestic poultry suggested that outbreaks most frequently took place in the four months during the cold season, including November, December, January, and February, while only a few outbreaks occurred during warm months (April, May, June, and September). Higher numbers of dead animals were identified in November, June, February, and January; during these months, the numbers of HPAI outbreaks were larger and more provinces were affected by HPAI. The most frequently identified type of HPAI viruses was HPAI H5N1; this strain was identified during every month, whereas HPAI H5N2

and H5N6 mainly occurred in January, February, April, August, September, and December. HPAI infections in chickens, ducks, and geese were identified during more than 8 months, and HPAI-infected wild birds were identified only in January, February, April, and May. It has been demonstrated that the spatial characteristics of HPAI outbreaks may be indicated by nationwide and regional outbreak distributions. The nationwide distribution in China showed that during 2004 to 2015, HPAI occurred in 26 provinces but not in Beijing, Shandong, Fujian, Chongqing, and Hainan. HPAI outbreaks appeared most frequently in Xinjiang, Hubei, and Guangdong, while larger numbers of dead animals were identified in Liaoning and Shanxi. HPAI H5N1 was identified in 25 provinces but not in Heilongjiang, while HPAI H5N2 mainly occurred in Hebei and Jiangsu, and HPAI H5N6 was detected in Heilongjiang, Jiangsu, Henan, and Guangdong. Among the species of infected animals, chickens were diagnosed with HPAI viruses in 23 provinces but not in Qinghai, Zhejiang, and Shanghai, while infected ducks and geese were identified in the southern provinces, and infected wild birds were identified in the northern provinces. In addition, HPAIinfected quail and rare birds were identified in Guizhou, Jiangsu, and Hunan provinces. The regional distribution of HPAI outbreaks in China suggested that HPAI outbreaks most frequently took place in the western region of China, followed by the central and eastern regions. However, the largest number of dead animals was identified in the eastern region of China, followed by the western and central regions. HPAI H5N1 occurred in the eastern, central, and western regions of China, while HPAI H5N2 mainly appeared in the eastern region of China, and HPAI H5N6 was identified in the eastern and central region of China. The species of infected animals did not differ substantially between different regions; however, chickens and wild birds were most frequently infected with in the western region, while infected ducks and geese were identified in the central and eastern regions, respectively. These results imply a need to put greater efforts into prevention and control strategies to avoid the possibility of continuous or larger HPAI outbreaks. In this sense, governments should therefore play a more critical role in this effort by taking several steps that include developing targeted surveillance and control policies toward various host outbreaks in different regions and during different time periods; furthermore, an effective vaccination strategy should be established for preventing and curbing the emergence of HPAI viruses. Additionally, the free-range and small-scale poultry production system should be replaced by large-scale farming, which has relatively high levels of biosecurity, hygiene, and disease prevention practices, all of which are important ways to reduce the number of HPAI outbreaks. Several aspect of the analysis presented in this study could be improved. The results of this study were based on the distribution of HPAI outbreak data that mostly came from the Official Veterinary Bulletin during 2004

CHARACTERISTICS OF HPAI OUTBREAKS

to 2015, which was published by the Chinese Ministry of Agriculture. However, there was a lack of information regarding HPAI virus infections in humans, which led to some limitations in this study. Additionally, this paper was based only on secondary data. This was one of the fundamental limitations of this study. Future researchers should focus on the collection of primary data from in-depth interviews of people who are involved in poultry farming and related stakeholders.

CONCLUSION In this study, we used the entire country of China as our study area and identified HPAI outbreaks from 2004 to 2015 to analyze the temporal and spatial characteristics of HPAI outbreaks; namely, we identified the frequency of HPAI outbreaks, numbers of dead animals, types of HPAI viruses, and species of infected animals. In this analysis, the interannual and intermonthly characteristics and nationwide and regional distributions contributed to the temporal and spatial characteristics of HPAI outbreaks, respectively. Our findings may be helpful to better understand the characteristics of HPAI outbreaks and distribution of HPAI outbreak risk, which can be used to improve the prevention and control of HPAI outbreaks in China.

ACKNOWLEDGEMENTS This study was supported by grant No. 14BJY121 from the National Social Science Foundation of China.

REFERENCES Capua, I., and D. J. Alexande. 2009. Avian influenza infection in birds: a challenge and opportunity for the poultry veterinarian. Poul. Sci. 88:842–846. Cattoli, G., I. Monne, A. Fusaro, T. M. Joannis, L. H. Lombin, M. M. Aly, A. S. Arafa, K. M. Sturm-Ramirez, E. Couacy-Hymann, J. A. Awuni, K. B. Batawui, K. A. Awoume, G. L. Aplogan, A. Sow, A. C. Ngangnou, I. M. El Nasri Hamza, D. Gamati´e, G. Dauphin, J. M. Domenech, and I. Capua. 2009. Highly pathogenic avian influenza virus subtype H5N1 in Africa: a comprehensive phylogenetic analysis and molecular characterization of isolates. PLoS One. 4:e4842. Fang, L. Q., S. J. de Vlas, S. Liang, C. W. Looman, P. Gong, B. Xu, L. Yan, H. Yang, J. H. Richardus, and W. C. Cao. 2008. Environmental factors contributing to the spread of H5N1 avian influenza in Mainland China. PLoS One. 3:e2268. FAO, 2013. Food and Agriculture Organization of the United Nations Statistics Division. Available: http://faostat3.fao.org/ browse/Q/QL/E. Farnsworth, M. L., and M. P. Ward, 2009. Identifying spatiotemporal patterns of transboundary disease spread: examples using avian influenza H5N1 outbreaks. Vet. Res. 40:336. Gao, R., B. Cao, Y. Hu, Z. Feng, D. Wang, W. Hu, J. Chen, Z. Jie, H. Qiu, K. Xu, X. Xu, H. Lu, W. Zhu, Z. Gao, N. Xiang, Y. Shen, Z. He, Y. Gu, Z. Zhang, Y. Yang, X. Zhao, L. Zhou, X. Li, S. Zou, Y. Zhang, X. Li, L. Yang, J. Guo, J. Dong, Q. Li, L. Dong, Y. Zhu, T. Bai, S. Wang, P. Hao, W. Yang, Y. Zhang, J. Han, H. Yu, D. Li, G. F. Gao, G. Wu, Y. Wang, Z. Yuan, and Y. Shu. 2013. Human infection with a novel avian-origin influenza A (H7N9) virus. N. Engl. J. Med. 368:1888–1897. Gilbert, M., X. M. Xiao, D. U. Pfeiffer, M. Epprecht, S. Boles, C. Czarnecki, P. Chaitaweesub, W. Kalpravidh, P. Q. Minh, M.

9

J. Otte, V. Martin, and J. Slingenbergh. 2008. Mapping H5N1 highly pathogenic avian influenza risk in Southeast Asia. Proc. Natl. Acad. Sci. U.S.A. 105:4769–4774. Huang, Z. Y., and J. M. Wang. 2015. Analysis of occurrence and characteristics of avian influenza in China during 2004–2014. Guangdong Agric. Sci. 42:93–98. (in Chinese). Jiang, W., S. Liu, G. Y. Hou, J. P. Li, Q. Zhuang, S. C. Wang, P. Zhang, and J. M. Chen. 2012. Chinese and global distribution of H9 subtype avian influenza viruses. PLoS One. 7:e52671. Lee, E. K., H. M. Kang, K. Kim, J. G. Choi, T. L. To, T. D. Nguyen, B. M. Song, J. Jeong, K. S. Choi, J. Y. Kim, H. S. Lee, Y. J. Lee, and J. H. Kim. 2015. Genetic evolution of H5 highly pathogenic avian influenza virus in domestic poultry in Vietnam between 2011 and 2013. Poult. Sci. 94:650–661. Li, R., Z. Jiang, and B. Xu. 2014. Global spatiotemporal and genetic footprint of the H5N1 avian influenza virus. Int. J. Health Geogr. 13:1–11. Li, X. L., K. Liu, H. W. Yao, Y. Sun, W. J. Chen, R. X. Sun, S. J. de Vlas, L. Q. Fang, and W. C. Cao. 2015. Highly pathogenic avian influenza H5N1 in Mainland China. Int. J. Environ. Res. Public Health. 12:5026–5045. Loth, L., M. Gilbertb, M. G. Osmanid, A. M. Kalama, and X. M. Xiao. 2010. Risk factors and clusters of highly pathogenic avian influenza H5N1 outbreaks in Bangladesh. Prev. Vet. Med. 96:104– 113. Martin, V., D. U. Pfeiffer, X. Y. Zhou, X. M. Xiao, D. J. Prosser, F. S. Guo, and M. Gilbert. 2011. Spatial distribution and risk factors of highly pathogenic avian influenza (HPAI) H5N1 in China. PLoS Pathog. 7:741–743. Oyana, T. J., D. Dajun, and K. E. Scott. 2006. Spatiotemporal distributions of reported cases of the avian influenza H5N1 (bird flu) in Southern China in early 2004. Avian Dis. 50:508–515. Pfeiffer, D. U., P. Q. Minh, V. Martin, M. Epprecht, and M. J. Otte. 2007. An analysis of the spatial and temporal patterns of highly pathogenic avian influenza occurrence in Vietnam using national surveillance data. Vet. J. 174:302–309. Si, Y., A. K. Skidmore, T. Wang, W. F. de Boer, P. Debba, A. G. Toxopeus, L. Li, and H. H. Prins. 2009. Spatio-temporal dynamics of global H5N1 outbreaks match bird migration patterns. Geospat. Health. 4:65–78. Si, Y. L., P. Debba, A. K. Skidmore, A. G. Toxopeus, and L. Li. 2008. Spatial and temporal patterns of global H5N1 outbreaks. ISPRS J. Photogramm. Remote Sens. 37(B2):69–74. Suarez, D. L. 2000. Evolution of avian influenza viruses. Vet. Microbiol. 74:15–27. Tian, H. Y., Y. Cui, L. Dong, S. Zhou, X. W. Li, S. Q. Huang, R. F. Yang, and B. Xu. 2015. Spatial, temporal and genetic dynamics of highly pathogenic avian influenza A (H5N1) virus in China. BMC Infect. Dis. 15:1–15. Tiensin, T., M. Nielen, T. Songserm, W. Kalpravidh, P. Chaitaweesub, A. Amonsin, S. Chotiprasatintara, A. Chaisingh, S. Damrongwatanapokin, S. Wongkasemjit, C. Antarasena, V. Songkitti, K. Chanachai, W. Thanapongtham, and J. A. Stegeman. 2007. Geographic and temporal distribution of highly pathogenic avian influenza: a virus (H5N1) in Thailand, 20042005: an overview. Avian Dis. 51:182–188. Tong, S. X., Y. Li, P. Rivailler, C. Conrardy, D. A. A. Castillo, L. M. Chen, S. Recuenco, J. A. Ellison, C. T. Davis, I. A. York, A. S. Turmelle, D. Moran, S. Rogers, M. Shi, Y. Tao, M. R. Weil, K. Tang, L. A. Rowe, S. Sammons, X. Y. Xu, M. Frace, K. A. Lindblade, N. J. Cox, L. J. Anderson, C. E. Rupprecht, and R. O. Donis. 2012. A distinct lineage of influenza: a virus from bats. Proc. Natl. Acad. Sci. U. S. A. 109:4269–4274. Wang, Y., Z. Jiang, Z. Jin, H. Tan, and B. Xu. 2013. Risk factors for infectious diseases in backyard poultry farms in the Poyang Lake area, China. PLoS One. 8:1475–1478. Zhang, R. J., E. J. Ge, S. F. Zhang, and M. Office. 2014. Analysis of global distribution of highly pathogenic avian influenza H5N1 with geographic information system. Chin. J. Public Health. 30:26–29. (in Chinese). Zhang, R. J., Y. D. Wei, and E. J. Ge. 2015. An analysis on spatial and temporal distribution of highly pathogenic avian influenza H5N1. Zhejiang J. Prev. Med. 27:653–656. (in Chinese).