Int J Biometeorol DOI 10.1007/s00484-013-0764-6

ORIGINAL PAPER

Short-term effects of meteorological factors on children hand, foot and mouth disease in Guangzhou, China Chun Chen & Hualiang Lin & Xiaoquan Li & Lingling Lang & Xincai Xiao & Peng Ding & Peng He & Ying Zhang & Ming Wang & Qiyong Liu

Received: 20 April 2013 / Revised: 2 October 2013 / Accepted: 4 November 2013 # ISB 2013

Abstract Hand, foot and mouth disease (HFMD) is a contagious viral illness that commonly affects infants and children. The underlying risk factors have not yet been systematically examined. This study analyzed the short-term effects of meteorological factors on children HFMD in Guangzhou, China. Daily count of HFMD among children younger than 15 years and meteorological variables from 2009 to 2011 were collected to construct the time series. A generalized additive model was applied to estimate the effects of meteorological factors on HFMD occurrence, after adjusting for long-term trend, seasonal trend, day of week, and public holidays. A negative association between temperature and children HFMD occurrence was observed at lag days 1–3, with the relative risk (RR)

Chun Chen, Hualiang Lin, and Xiaoquan Li contributed to this work equally. C. Chen : X. Li : X. Xiao : P. Ding : P. He : Y. Zhang : M. Wang (*) Guangzhou Center for Disease Control and Prevention, Guangzhou, China e-mail: [email protected] H. Lin : L. Lang Guangdong Provincial Institute of Public Health, Guangdong Provincial Center for Disease Control and Prevention, Guangzhou, China Q. Liu (*) Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases; State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Control and Prevention, China CDC, Beijing, China e-mail: [email protected] Q. Liu Shandong University Climate Change and Health Center, Jinan, China

for a 1 °C increase on lag day 2 being 0.983 (95 % confidence intervals (CI) 0.977 to 0.989); positive effect was found for temperature at lag days 5–9, with the highest effect at lag day 6 (RR=1.014, 95 % CI 1.006 to 1.023). Higher humidity was associated with increased HFMD at lag days 3–10, with the highest effect at lag day 8 (RR=1.009 for 1 % increase in relative humidity, 95 % CI 1.007 to 1.010). And we also observed significant positive effect for rainfall at lag days 4 and 8 (RR=1.001, 95 % CI 1.000 to 1.002) for 1-mm increase. Subgroup analyses showed that the positive effects of temperature were more pronounced among younger children. This study suggests that meteorological factors might be important predictors of children HFMD occurrence in Guangzhou. Keywords Hand . Foot and mouth disease . Weather . Mean temperature . Relative humidity . Guangzhou

Introduction Hand, foot and mouth disease (HFMD) is an emerging viral infectious disease that mainly affects infants and children, the main symptom includes fever, mouth ulcers, and vesicles on the hands, feet, and mouth (Jiang et al. 2012). In most cases, this illness is mild and self-limiting, but more severe clinical symptoms may occur when there are complications, such as encephalitis, aseptic meningitis, and acute flaccid paralysis (Chen et al. 2007). HFMD is most frequently caused by Coxsackie virus A16 (CA16) and enterovirus 71 (EV71) (Zhang et al. 2009; Wu et al. 2010), among which EV71 is more commonly associated with severe symptoms, including central nervous system disorders and even deaths resulting from pulmonary edema in a small proportion of children, particularly those aged 5 years and younger (Zhang et al.

Int J Biometeorol

2010; Onozuka and Hashizume 2011). There has been no specific drug or vaccine available for HFMD so far, so preventive measures remain the only effective way to prevent its transmission, such as avoiding direct contact with infective persons, disinfection of viral contaminated environment, and good personal hygiene habits (Hii et al. 2011). It is projected that HFMD incidence will increase in the future due to continued viral mutation, climate change, and the lacking of health resources and effective surveillance systems in some countries (Hosoya et al. 2006; Huang et al. 2013). Early warning system that combines with information of disease surveillance and risk factors may also be helpful for the health authorities to issue potential disease outbreak information for the target population (Ma et al. 2010a, b). Asia-Pacific countries have experienced an increasing trend of HFMD outbreaks in the past decades, resulting in thousands of deaths among children. For example, a total of 1, 619,706 new HFMD cases were reported in China in the year of 2011, resulting in 509 deaths (Jin et al. 2012). Severe outbreaks have also been witnessed in Taiwan in 1998 with 129,106 HFMD cases and 78 children deaths being reported (Ho et al. 1999). Significant seasonality in the incidence of HFMD has been exhibited in a number of countries. For example, March–July peak was observed in Jiangsu Province, China (Mao et al. 2010). A bimodal seasonal pattern was reported in UK with peaks in summer and late autumn/early winter (Bendig and Fleming 1996), while the incidence of HFMD was highest in summer in Taiwan and Japan (Chen et al. 2007; Onozuka and Hashizume 2011). In Finland, most HFMD cases were reported in autumn (Blomqvist et al. 2010). The seasonality of HFMD incidence indicates that its incidence may be influenced by meteorological factors. However, the impact of weather variability on the incidence of HFMD has only been quantitatively investigated in a limited number of studies with inconsistent findings. For example, in Hong Kong, it has been suggested that the changing epidemiology (a new peak in winter) might be due to winter temperature increase (Ma et al. 2010a, b). In Japan, a time series analysis found that ambient temperature and relative humidity were associated with increased HFMD occurrence (Onozuka and Hashizume 2011). The study in Singapore showed that maximum temperature above 32 °C was linked with increased HFMD incidence (Hii et al. 2011), and a recent analysis suggested that the incidence of HFMD in Shenzhen might be influenced by El Niño phenomenon (Lin et al. 2013c). It was reasonable to hypothesize that weather variation can affect the transmission dynamic of HFMD. The current study examined the relationship of meteorological factors with the occurrence of children HFMD using surveillance data collected in Guangzhou, China, from 2009 to 2011. We also examined whether this association varied by age and sex.

Methods Setting Guangzhou is the capital city of Guangdong Province in southern China, about 180 km north of Hong Kong (Liu et al. 2013). It has a typical monsoon-influenced climate with wet and hot summers and dry and cool to mild winters (Lu et al. 2009). Guangzhou has a population of 12 million. The residents of four districts in urban Guangzhou (Yue Xiu, Li Wan, Hai Zhu, and Tian He) were selected as the subjects of the current study, which have an area of 280 km2 and are home to 5.1 million residents. These four districts were chosen for two reasons. Firstly, there are weather monitoring stations in these districts. Secondly, the HFMD data in these districts are of high quality. Figure 1 showed the geographical location of the study area in China. Data sources District-specific daily data on count of HFMD in children younger than 15 years old covering the period 2009–2011 were obtained from Guangzhou Center for Disease Control and Prevention. HFMD case was diagnosed based on clinical symptoms, which included vesicular lesions on hands, feet, mouth (which were often ulcerated), and, frequently, buttocks, in accordance with the National Guideline on diagnosis and treatment of Hand, Foot and Mouth Disease (Chinese Ministry of Health). According to China’s notifiable infectious disease regulations, all HFMD cases should be reported to the infectious disease surveillance system. A recent study has demonstrated that the reporting data are of good quality, especially in the eastern regions of China, with reporting completeness of 99.84 % and accuracy of the information reported of 92.76 % (Ji et al. 2011). We chose the children younger than 15 years old for the subjects as this illness mainly affected children of these age groups, for example, about 99 % of the cases were among children younger than 15 years in the study area. Meteorological data including daily mean temperature, relative humidity, wind velocity, and rainfall were obtained from Guangdong Provincial Weather Data Sharing System (http://www.grmc.gov.cn/), which was publicly accessible. Statistical analysis As daily count of HFMD typically follows a Poisson distribution, we used generalized additive modeling (GAM) with a log link and allowing Poisson auto-regression and overdispersion to examine the relationship between daily meteorological variables and daily HFMD (Villeneuve et al. 2003). All models were adjusted for the day of week (DOW) and public holidays using categorical indicator variables. In

Int J Biometeorol Fig. 1 Geographical location of study area in China

addition, we used penalized smoothing splines (Kan et al. 2007; Lin et al. 2013a) to adjust for long-term trend and seasonal pattern in daily morbidity with degree of freedom (df) selected a priori based on previous studies (Bell et al. 2009; Peng et al. 2009). For the smooth functions of calendar time, 6 df per year was chosen so that we could filter out the information at time scales of 2 months. The model for temperature can be specified as:  .  log½EðYt Þ ¼ a þ temperature þ s t; df ¼ 6 year þ β 1 *DOW þ β2 *PH þ weather where E(Yt) is the expected number of HFMD cases on day t, α is the intercept, s() indicates a smoother based on penalized splines, df is the degree of freedom, DOW is an indicator for day of week, PH presents a binary variable for public holidays, and β is the regression coefficient, weather is the other weather variables, including relative humidity, wind velocity, rainfall. We estimated the linear effect of various weather variables according to different lag structures, including current day (lag 0) up to 10 days before (lag 10) as this infection typically has an incubation period of 3–7 days (Wong et al. 2010). District-

specific risk estimates were firstly evaluated, and combined effects were evaluated using a random effect meta-analysis model (Lin et al. 2013b). The robustness of the key findings was assessed by varying degree of freedom in the smooth function of time used to adjust for seasonal and long-term trends. To justify the assumption of linearity between the logarithm of HFMD count and the key weather variables (significantly related to HFMD incidence), we graphically examined concentration–response relationships derived using a smoothing function (Hong et al. 2002; Kan et al. 2003). We tested for effect modification of the significant association by age and sex with stratified analyses. We stratified the analyses by age group (≤2 years, 2–6 years, and 6–15 years of age), and sex (male and female). We tested the statistical significance of differences between the effect estimates of the strata divided by potential effect modifiers by calculating the 95 % confidence interval as. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðQ1 −Q2 Þþ1:96 ðSE1 Þ2 þ ðSE2 Þ2 where Q 1 and Q 2 were the effect estimates for each stratum (such as male and female), and SE1 and SE2 were their corresponding standard errors (Zeka et al. 2006).

Int J Biometeorol

All analyses were carried out using the “mgcv” package in the software R2.15.1 (R Development Core Team, 2012) (Wood 2006). We report results as relative risk (RR) with 95 % confidence intervals (CI) in daily HFMD count for one unit increase in each weather variables.

Results Between 1 January 2009 and 31 December 2011, there were 34,527 HFMD cases among children younger than 15 years of age being reported in the four districts in Guangzhou. During the study period, there were more male cases with a male-tofemale sex ratio of 1.75:1 (21,989:12,538). A total of 13,117 cases (38.0 %) were reported in infants aged below 2 years, 19,899 (57.6 %) were preschool children aged 2–6 years, 1, 511 (4.4 %) were those aged at 6–15 years. The descriptive summary for daily HFMD and weather conditions in the four districts was shown in Table 1. On average, there were 5.4, 5.2, 11.7, and 9.2 daily HFMD cases in Yue Xiu, Li Wan, Tian He, and Hai Zhu over the study period, respectively. Generally, similar weather conditions were observed in the four districts. In the four districts, the daily mean temperature were

Table 1 Distribution of daily data on children HFMD and weather in the four districts of Guangzhou (2009–2011)

Variable Yue Xiu District Daily HFMD cases Temperature (°C) Relative humidity (%) Rainfall (mm) Wind velocity (m/s) Li Wan District Daily HFMD cases Temperature (°C) Relative humidity (%) Rainfall (mm) Wind velocity (m/s) Tian He District Daily HFMD cases Temperature (°C) Relative humidity (%) Rainfall (mm) Wind velocity (m/s) Hai Zhu District Daily HFMD cases Temperature (°C) Relative humidity (%) Rainfall (mm) Wind velocity (m/s)

around 22 °C (ranging from 21.9 to 22.9 °C), daily relative humidity ranged from 71.6 to 75.1 %, rainfall was 4.2 to 5.0 mm, and wind velocity ranged from 1.6 to 2.1 m/s, respectively. Figure 2 showed the time series of daily HFMD cases and meteorological variables in the four districts during the study period. There were seasonal patterns in the meteorological factors. Summer epidemic peaks were observed with the highest in 2011. Figure 3 showed that the associations between various meteorological variables and HFMD for each district. For temperature, an adverse effect was generally observed at shorter lag days (0–3 days), and positive effect at longer lag days. Positive associations were observed for relative humidity in all the four districts and rainfall in Tian He District, whereas no significant association was found for wind velocity. Figure 4 illustrated the overall relative risk estimates for the key meteorological factors (with significant associations with HFMD) at different lag days estimated by a random effect meta-analysis approach. Negative association between temperature and HFMD at lag days 1–3 were observed, with the RR being 0.983 for 1 °C increase (95 % CI 0.977 to 0.989) at lag day 2; positive effect was found for temperature at lag days

Mean±SD

Min.

P(25)

P(50)

P(75)

Max.

5.4±5.7 22.3±6.4 72.2±13.0 5.0±15.3 1.6±0.9

0.0 5.5 27.0 0.0 0.1

1.0 17.8 65.0 0.0 1.0

4.0 24.1 74.0 0.0 1.4

8.0 27.7 83.0 1.8 1.9

41.0 32.3 99.0 214.7 6.8

5.2±5.0 22.3±6.4

0.0 5.5

1.0 17.9

4.0 23.9

7.0 27.6

30.0 32.3

72.2±13.0 5.0±15.3 1.7±0.9

27.0 0.0 0.1

65.0 0.0 1.0

74.0 0.0 1.4

82.0 1.8 1.9

99.0 214.7 6.9

11.7±12.1 21.9±6.2 75.1±10.4 5.0±14.7 2.1±0.8

0.0 6.0 32.0 0.0 0.6

3.0 17.3 70.0 0.0 1.5

8.0 23.4 76.0 0.0 1.9

16.0 27.1 82.0 1.9 2.5

71.0 31.2 96.0 180.8 5.9

9.2±9.7 22.9±6.3 71.6±12.3 4.2±13.0 2.0±0.8

0.0 6.1 26.0 0.0 0.4

2.0 18.2 65.0 0.0 1.4

5.0 24.4 72.0 0.0 1.8

14.0 28.1 81.0 1.6 2.4

53.0 32.6 95.0 165.0 6.0

Int J Biometeorol

Fig. 2 The time series of daily meteorological variables in four districts of Guangzhou, 2009–2011 (unit: temperature (degrees Celsius), relative humidity (percent), rainfall (millimeters), and wind speed (meters per second))

5–9, with the highest effect at lag day 6 (RR=1.014 for 1 °C increase, 95 % CI 1.006 to 1.023). Higher humidity was associated with increasing HFMD at lag days 3–10, with the highest effect being observed at lag day 8 (RR=1.009 for 1 % increase in daily relative humidity, 95 % CI 1.007 to 1.010). Consistent with the effect of relative humidity, we also observed significant positive effect for rainfall at lag days 4 and 8, with similar effect at 2 days (RR=1.001, 95 % CI 1.000 to 1.002) for 1-mm increase in daily rainfall. Table 2 illustrated the results of stratified analyses for temperature, relative humidity, and rainfall by sex and age group. We found statistically significant difference in the effect of temperature on lag day 6 among the age groups with higher effects on age groups of 0–2 years old and 2–6 years old, and nonsignificant effect among those aged 6–15 years were observed. No significant difference was observed for other comparisons. Some examples of the dose–response relationship for temperature, relative humidity, and rainfall with HFMD occurrence in the four districts were shown in Fig. 5. An approximately linear relationship was observed for daily temperature, relative humidity, and rainfall in the districts.

In the sensitivity analysis, we changed df (4, 5, 7, 8) for calendar time to control for seasonality and long-term trend, which yielded similar results.

Discussion This was one of the few studies examining the effects of meteorological factors on adolescent HFMD occurrence at a daily scale. The finding showed that daily mean temperature, relative humidity, and rainfall were related to children hand, foot and mouth disease in four districts of Guangzhou, suggesting that meteorological factors might play an important role in the transmission of HFMD in the study region. This study recommended that these meteorological factors should be considered in the early warning system for this infection. The finding of the association between temperature, relative humidity, rainfall, and HFMD in the current study was generally consistent with findings in different regions of the world (Ma et al. 2010a, 2010b; Onozuka and Hashizume 2011; Wang et al. 2011). Several postulations might be helpful to better understand the possible pathways by which

Int J Biometeorol

Fig. 3 Relative risk (RR with 95 % CI) in hand, foot and mouth disease for one unit increase in daily meteorological factors for the current day (lag 0) to 10 days before the current day (lag 10) in four districts of Guangzhou

Fig. 4 Meta-analysis of relative risk (RR with 95 % CI) in hand, foot and mouth disease for one unit increase in daily meteorological factors for the current day (lag 0) to 10 days before the current day (lag 10)

Int J Biometeorol Table 2 Relative risk (RR with 95 % CI) of temperature and relative humidity associated children HFMD, stratified by sex and age group

Variable

RR (95 % CI) Temperature (lag 2)

Temperature (lag 6)

Humidity (lag 8)

Rainfall (lag 8)

0.984 (0.975–0.993) 0.974 (0.964–0.984)

1.016 (1.007–1.025) 1.011 (1.001–1.021)

1.009 (1.007–1.011) 1.009 (1.006–1.011)

1.001 (1.000–1.002) 1.001 (1.000–1.002)

0.986 (0.976–0.996) 0.977 (0.967–0.986) 0.983 (0.962–1.000)

1.010 (1.000–1.021) 1.018 (1.008–1.028) 0.991 (0.970–1.010)*

1.009 (1.006–1.011) 1.009 (1.007–1.012) 1.004 (0.999–1.010)

1.001 (1.000–1.002) 1.001 (1.000–1.003) 1.002 (0.999–1.003)

Sex

*p

Short-term effects of meteorological factors on children hand, foot and mouth disease in Guangzhou, China.

Hand, foot and mouth disease (HFMD) is a contagious viral illness that commonly affects infants and children. The underlying risk factors have not yet...
614KB Sizes 0 Downloads 0 Views