1 Introduction

Hand, foot, and mouth disease (HFMD) is a relatively common syndrome, which mainly occurs in children under 5 year-old [1], and is usually caused by intestinal viruses of the Picornaviridae family including human enterovirus 71 (EV 71) [2] and coxsackievirus (CoxA 16) [3]. EV 71 was first reported in New Zealand in 1957, and the CoxA 16 was first isolated in Canada in 1958, and now has a worldwide distribution. HFMD outbreaks associated EV 71 were reported in Australia in 1972 [4], Japan in 1973 [5], Bulgaria in 1975, and Hungary in 1978. In 1997, a serious outbreak in Sarawak killed 29 people [6] and this was followed by a series of outbreaks in the Asia Pacific region [7–9]. Although many HFMD-infected adults remain asymptomatic, cases in children start with flu-like symptoms that can rapidly develop to fever, headache [10], vomiting, and sore throat [11], accompanied by vesicles and ulcers on the gums, tongue, buccal mucosa, and palate [10–12]. Infection can lead to meningitis and swelling of the brain, with the possibility of long-term neurological damage or even death [11, 12]. The disease can be transmitted directly from person to person by contact with nasal discharge, saliva, feces, and fluid from blisters of infected persons, and no animal reservoir is necessary. Additionally, it can be spread through food and water infected by intestinal viruses [10–12]. Unfortunately, as there is no specific treatment or vaccine for the disease, HFMD still remains a significant public health issue.

In mainland China, the first HFMD case was reported in 1981 in Shanghai [13], followed by cases in Beijing, Hebei, Tianjin, and Fujian, and it eventually occurred in almost all provinces and regions of China [14]. The disease caused serious outbreaks in Fuyang City in March 2008 and more than 500,000 cases were reported in mainland China [15], of whom about 200 children died. Over the past few years, some studies on the epidemiological characteristics of HFMD have been carried out in China [16–18]. However, little is known about dynamic of HFMD transmission in Changchun. Between 2008 and 2011, over 17,000 cases of HFMD were reported in Changchun, northeastern China. Thus in this study, we aimed to characterize the temporal and spatial patterns of HFMD in Changchun, and to identify the risk factors for occurrence of HFMD at township level.

2 Methods

2.1 Study area

The study area encompassed Changchun, the capital city of Jilin Province, northeastern China, located between latitude 43.26° to 45.3° north, and longitude 124.5° to 127.2° east (Fig. 1). Changchun includes 119 townships in 10 counties with a total land area of 20,660 km2 and about 7.7 million inhabitants.

Fig. 1
figure 1

Map of the study area, annual incidence of hand, foot and mouth disease (HFMD) in each township and main transportation lines in Changchun

Full size image

2.2 Data collection and management

2.2.1 Clinical definitions

HFMD was classified as a class C notifiable disease by the Ministry of Health on May 2, 2008. According to the Law for Prevention and Control of Infectious Disease in China, information must be submitted to the web-based China Information System for Disease Control and Prevention (CISDCP) within 24 h of a diagnosis based on clinical symptoms according to the diagnostic code in the guidebook published by the Ministry of Health, China [19]. A reported case was defined as a person with fever and accompanying skin papular/vesicular erythema on the hands, feet, mouth, and buttocks, though some cases may not have fever [17, 18]. In 2008, a severe case was defined as a person with additional manifestations such as myoclonus, acute flaccid paralysis, encephalitis, cardiopulmonary failure, or pulmonary edema. Since 2009, a case was deemed severe if it also included respiratory or circulatory complications, plus increased peripheral white blood cells, abnormal cerebrospinal fluid, increased blood sugar, or any abnormality in an electroencephalogram, cerebrospinal magnetic resonance image, chest X-ray, or ultrasound cardiogram. A few cases may not have a rash and laboratory analysis has been mandatory since 2009 [17]. Clinical specimens, e.g., from the pharynx, vesicle or feces, of at least 10 patients each month from March to October each year were collected and tested for EV 71, CoxA 16, and other enterovirus by reverse transcription polymerase chain reaction (RT-PCR). The specimens are obtained from all severe and fatal cases and some other reported cases [17]. Laboratory tests were conducted by Changchun Center for Diseases Control and Prevention (Changchun CDC) according to the guidebook published by the Ministry of Health, China.

2.2.2 Data collection

Data on reported cases, severe cases, and tested cases from January 2008 to December 2011 were obtained from CISDCP, which included information on sex, age, residential address, date of onset of symptoms, and occupation. The 2010 basic census information was obtained from the Bureau of Statistics of Changchun. A digital township-level map of Changchun was obtained from the Data Sharing Infrastructure of Earth System Science (http://www.geodata.cn/). Climate data (mean temperature, relative humidity) were obtained from the Meteorological Bureau of China and managed to match the onset date of the disease. GDP raster was obtained from the Data Sharing Infrastructure of Earth System Science (http://www.geodata.cn/), and was extracted by the boundary of study area then calculated at township-level. Data of transportation routes were extracted from maps of China’s transportation routes including freeway and national highway; the nearest distance from each township’s centroid to transportation lines (freeway and national highway) was calculated by proximity tools in ArcGIS (ESRI Inc., Redlands, CA, USA). Data on kindergartens and medical facilities were obtained from Changchun CDC and geo-referenced to the map of Changchun according to their locations. In addition, all townships were classified into three categories: urban, urban–rural border, and rural. Urban is identified as the township containing the government of county, urban- rurally border is identified as a township having one or more conjunct boundaries with an urban township, and others are classified into rural townships.

2.3 Analysis of the epidemiological distribution of HFMD

To display the temporal distribution of HFMD in Changchun, epidemic curves were created by plotting the weekly number of newly reported cases and severe cases, as well as the weekly mean temperature and relative humidity. The incidences for each sex over age groups were calculated using 2010 census data by using the cumulative cases over age groups in the study period. Each reported case was geo-referenced to the map of Changchun according to its address at the level of township, and the incidence for each township was calculated by using cases under 5 years old as the numerator, and the susceptible population (all children under 5 years old) as the denominator. The annual incidence was mapped at the township level using ArcGIS 9.3 software (ESRI Inc., Redlands, CA, USA). In addition, main transportation lines such as national highways, freeways were overlapped on the annual incidence map (Fig. 1).

2.4 Statistical analysis

To explore the association between HFMD incidence and potential influencing factors, Poisson regression was performed. The cumulative number of HFMD cases in children under 5 years old in each township was set as the outcome variable, and the number of children under 5 years old in each county was included as the offset. Potential environmental and socio-economic factors, such as population density, children population density, type of township, distance to the nearest national highway and to the nearest freeway, GDP, density of medical facilities, and density of kindergartens for each township were included as co-variables in the analysis. The percentage change (PC) in incidence in response to the change of a variable by a given amount (which is equal to \( 100*\left( {{ \exp }\left( {\text{coefficient}} \right) - 1} \right) \)), (1,000 persons per km2 for population density, 100 children per km2 for child density, 10 km for the distance to the nearest national highway and to the nearest freeway, 1000 RMB per capita for GDP, one medical facility per 10,000 persons for the density of medical facilities and one kindergarten per 10,000 persons for the density of kindergartens), was used to determine the impact of each variable on HFMD incidence. The 95 % confidence intervals (CI) and corresponding P values were estimated after correcting for over dispersion because of the nature of infectious diseases with spatial clustering patterns [20]. A univariate analysis was first performed to examine the effect of each variable separately, followed by a multivariate analysis, which include all co-variables with a P value < 0.2 in the univariate analysis. The analysis was performed using STATA 9.1 software (StataCorp LP, College Station TX, USA).

3 Results

A total of 17,464 HFMD cases were reported in Changchun from 2008 to 2011, including 304 severe cases and eight deaths. Figure 2a shows the epidemic curves of the weekly number of newly reported cases. The annual recurrence started at about week 16 (end of April), peaked at about week 30 (middle of August), and then decreased during the next 6–7 weeks. The weekly number of reported HFMD cases were positively correlated with both the mean temperature (Spearman’s rho = 0.786, P < 0.001) and relative humidity (Spearman’s rho = 0.271, P < 0.001). The cumulative incidence for different age groups varied and it peaked at the 2 years old about 463 per 10,000 children. In addition, 86.8 % of cases were under 5 years old, 65.4 % of cases were pre-school children who stayed at home, and 29.1 % of cases were pre-school children attending kindergarten. The highest incidence occurred at 2 years of age in both boys and girls, and there was a higher incidence in boys than in girls (χ 2 = 130.03, P < 0.001) (Fig. 3). Figure 2b shows the epidemic curves of weekly number of newly severe cases each year. A total of 304 severe cases were reported during the study periods including 113 girls and 191 boys. Among all reported cases, the gender ration between boys and girls in severe cases was 1.69:1 and 1:48:1 in mild cases. The age distribution of severe cases was younger than the mild cases (F = 5.4, P < 0.001). About 62.5 % mild cases were pre-school children who stayed at home and about 73.7 % severe cases were pre-school children who stayed at home. In this study, a total of 895 cases underwent viral testing, and viral nucleic acid was detected in 73.1 % (654/895) (Fig. 4), including 369 EV 71, 135 CoxA 16, and 150 other enterovirus. There was no co-infection of EV71 and CoxA 16, and a significantly higher detected rate was found for EV71 infection (χ 2 = 151.2, P < 0.001) relative to CoxA 16 infection. Figure 1 shows the spatial distribution of the annual incidence of HFMD at the township level. The disease spanned the whole area, with the annual incidence ranging from 9 to 1,169 per 10,000 children (<5 years old), with a mean of 141 cases per 10,000 children. The cluster of townships with a high incidence were mainly located in the central–southern Changchun, the central urban areas, and a higher incidence were found in urban areas and urban–rural border areas with a median of 242 cases and 135 cases, respectively, per 10,000 persons compared with rural areas(75 per 10,000 persons).

Fig. 2
figure 2

Epidemic curves of HFMD and weekly climate variables (2008–2011) in Changchun, northeastern China. a The weekly temporal distribution of reported cases. b The weekly temporal distribution of severe HFMD cases. c The weekly temporal distribution of mean temperature. d The weekly temporal distribution of relative humidity

Full size image
Fig. 3
figure 3

Cumulative incidence for boys and girls for different ages

Full size image
Fig. 4
figure 4

Results of laboratory viral testing of 895 cases

Full size image

Univariate Poisson regression analyses revealed that HFMD incidence was significantly associated with type of township (urban area, urban–rural border area, or rural area), distance to the nearest national highway, distance to the nearest freeway, GDP per capita, density of medical facilities, and density of kindergartens (Table 1). Multivariate analysis showed that occurrence of HFMD was significantly associated with type of township, distance to the nearest freeway, and GDP per capita, while the significant effect of distance to the nearest national highway, density of medical facilities, and density of kindergartens disappeared after correction for adjustment other factors (Table 1).

Table 1 Association between incidence of HFMD and risk factors at the township-level, Changchun, Northeastern China, 2008–2011, by Poisson regression
Full size table

4 Discussion

HFMD remains an important public health challenge in China. After serious outbreaks, HFMD was included as a class C notifiable disease in China from May 2008 [17, 18]. Our study provides an overview of the epidemiological characteristics of HFMD in Changchun. Pre-school children under 5 years old of age accounted for 86.8 % of all cases. Children under 1 year had a relatively low incidence, possibility because of a lower chance of exposure and maternal antibodies [21]. A significant difference in the incidence among different age groups and sex groups (<6 years) was observed in this study. Boys had a higher risk of infection, which may be related to the behavior of boys, who may be more restless and badly behaved, allowing more opportunities to contract the viruses [22, 23]. After the serious outbreaks of HFMD in China, the government paid more attention to intervention measures for children in nursery schools, e.g., measurement of body temperature, examination of popular/vesicular erythema on hands, mouth each morning before school, and opening windows to improve the environment, but there remained a lack of suitable measures for pre-school children staying at home [18]. In the past decade, EV 71 has established itself as a regular cause of HFMD outbreaks around the pacific region. This study found that varied virus-related HFMD infections existed in Changchun, but that EV 71 infection played a primary role, remained a different picture with the epidemic in Germany [24], where CoxA 16 was more common. A seasonal variation in HFMD was observed in Changchun, as was also observed in Singapore [25], Sarawak, and Japan [26] by previous studies. However, the peak time was one week earlier than that in Naganno [27], suggesting the existences of geographic variations. Correlation between the incidence of HFMD and climate variables has been reported in Singapore, Sarawak, and Japan. Our study found significant associations between HFMD epidemic and mean temperature and relative humidity in Changchun [17, 25, 26]. We can conclude that these climate variables can influence the dynamics of the disease. Higher temperature and relative humidity could aid the transmission of the disease [17, 25, 26].

A higher incidence of HFMD in Changchun was observed in urban areas and urban–rural border areas than in rural areas; differ from the studies in Harbin, Huizhou, and Tokyo [18, 23, 26]. Transportation routes have been recognized as an importation role in accelerating the spread of severe acute respiratory syndrome and pandemic influenza [28, 29], our study showed occurrence of HFMD was negative correlative to the distance to the freeway, which indicates that local transmission could contribute to the spread of the disease. Migration of humans along those transportation lines will spread the disease in geographic areas and population [29]. The type of township (urban, urban–rural border areas) posed a high risk for occurrence of the disease, possibly related to the frequent movement of citizens in these areas, thus directly increasing the risk of HFMD transmission from human to human [28, 29].The urban and urban–rural border areas with a much more GDP per capita posed a high risk for the occurrence of the disease, which were accompanying frequently migration of human, which directly increase the chances of HFMD transmission between human and human [28, 29]. In addition, the association between HFMD incidence and density of medical facilities was excluded after adjustment for other factors, which indicates that CISDCP could not influence the reporting of HFMD cases [29].

In conclusion, this study found that the incidence of HFMD in Changchun, 2008–2011, showed a seasonal variation and was highest in children under the age of 5 years, with more boys than girls infected. It was higher in urban and urban–rural border areas than in rural areas and was associated with distance to the nearest freeway and GDP per capita. There are some limitations, for example, the study period is not long enough. Future studies may focus on those points including circulating virus serotypes, intervention measures for pre-school children staying at home, and personal risk behavior, which could contribute to the temporal and spatial distribution of HFMD.