Int J Biometeorol DOI 10.1007/s00484-015-1023-9
ORIGINAL PAPER
The association between diurnal temperature range and childhood bacillary dysentery Li-ying Wen 1 & Ke-fu Zhao 2 & Jian Cheng 1 & Xu Wang 1 & Hui-hui Yang 1 & Ke-sheng Li 1 & Zhi-wei Xu 3,4 & Hong Su 1
Received: 12 June 2014 / Revised: 22 May 2015 / Accepted: 26 May 2015 # ISB 2015
Abstract Previous studies have found that mean, maximum, and minimum temperatures were associated with bacillary dysentery (BD). However, little is known about whether the within-day variation of temperature has any impact on bacillary dysentery. The current study aimed to identify the relationship between diurnal temperature range (DTR) and BD in Hefei, China. Daily data on BD counts among children aged 0–14 years from 1 January 2006 to 31 December 2012 were retrieved from Hefei Center for Disease Control and Prevention. Daily data on ambient temperature and relative humidity covering the same period were collected from the Hefei Bureau of Meteorology. A Poisson generalized linear regression model combined with a distributed lag non-linear model (DLNM) was used in the analysis after controlling the effects of season, long-term trends, mean temperature, and relative humidity. The results showed that there existed a statistically Li-ying Wen and Ke-fu Zhao contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s00484-015-1023-9) contains supplementary material, which is available to authorized users. * Hong Su [email protected] 1
Department of Epidemiology and Health Statistics, School of Public Health, Anhui Medical University, 81 Meishan Road, Hefei 230032, Anhui, China
2
Hefei Center for Disease Control and Prevention of Anhui Province, Hefei 230061, Anhui, China
3
School of Public Health and Social Work, Queensland University of Technology, Kelvin Grove, Brisbane, QLD 4509, Australia
4
Institute of Health and Biomedical Innovation, Queensland University of Technology, Kelvin Grove, Brisbane, QLD 4509, Australia
significant relationship between DTR and childhood BD. The DTR effect on childhood bacillary dysentery increased when DTR was over 8 °C. And it was greatest at 1-day lag, with an 8 % (95 % CI=2.9–13.4 %) increase of BD cases per 5 °C increment of DTR. Male children and children aged 0–5 years appeared to be more vulnerable to the DTR effect. The data indicate that large DTR may increase the incidence of childhood BD. Caregivers and health practitioners should be made aware of the potential threat posed by large DTR. Therefore, DTR should be taken into consideration when making targeted health policies and programs to protect children from being harmed by climate impacts. Keywords Climate change . Diurnal temperature range . Childhood . Bacillary dysentry . Hefei
Introduction Bacillary dysentery (BD), a bacterial infection of the mucosal surface of the intestines caused by different species of Shigella bacteria, including Shigella dysenteriae, Shigella flexneri, Shigella boydii, and Shigella sonnei (Niyogi. 2005), often results in severe diarrhea. Patients who are infected with Shigella frequently develop bloody diarrhea, fever, abdominal cramps, and tenesmus with an incubation period of 1 or 2 days (Niyogi 2005). The transmission of BD usually occurs via polluted drinking water and contaminated food or through person-toperson contact. It is a public health issue in both developed and developing countries (Zhang et al. 2008), particularly in less developed countries (Kotloff et al. 1999; von Seidlein et al. 2006; Walker et al. 2012; Wang et al. 2005, 2006). Currently, no specific drug or targeted vaccine is available for BD (Phalipon et al. 2008). But many researchers have been trying to develop a vaccine even in the face of many
Int J Biometeorol
challenges and difficulties (Barry et al. 2013; Camacho et al. 2013; Kim et al. 2013). In particular, the Chinese government has made great efforts in the prevention and control of BD, and has made obvious progresses. According to the National Notifiable Disease of China, the reported BD cases decreased from 205,972 in 2012 to 187,303 in 2013 (Ministry of Health, P.R. China.). Although the incidence of BD has declined largely, a considerable disease burden still exists (Zhang et al. 2012), especially among children (Barry et al. 2013; Ghaemi et al. 2007; Wang et al. 2005, 2006). Thus, it can be seen that BD in children would attract increasing attention. A variety of determinants, including environmental, nutritional, socioeconomic, and behavioral factors, can promote the occurrence of BD (Boru et al. 2013; Haley et al. 2010). Actually, in recent years, the meteorological factors have caused more and more research attention (Li et al. 2013a, b; Luo et al. 2010) and mean, maximum, and minimum temperatures have been associated with the occurrence of BD (Gao et al. 2014; Guan et al. 2008; Li et al. 2013a, b; Zhang et al. 2007, 2008). Amazingly, another meteorological indicator, e.g., diurnal temperature range (DTR), refined as the difference between maximal and minimal temperatures within 1 day, has sparked extraordinary interests. The reason may be that DTR can incorporate the information of maximum and minimum temperatures and reflect whether the weather is stable or not (Kalnay and Cai 2003). Previous researches have identified that DTR is correlated with respiratory and cardiovascular diseases (Chen et al. 2007; Lim et al. 2012a, b; Song et al. 2008; Xu et al. 2013a). However, the influence of temperature variation on BD has received little attention so far. Additionally, few studies have investigated the effect of DTR on children’s health (Xu et al. 2013b), and little is known about the relationship between DTR and childhood BD. In the light of the above considerations, we hypothesized that DTR was correlated with the incidence of childhood BD. This study emphasized four key research issues: (1) What is the relationship between DTR and BD? (2) Which children subgroups are most prone to DTR effect? (3) Is there any difference of DTR effect on childhood BD between two genders? (4) Is there a delayed effect of DTR on childhood BD?
Materials and methods Data collection Hefei is the capital city of Anhui province, located on the east-central of China (31° 52 N, 117° 17 E). It is compromised of nine counties, with a total land area of about 11,408 km2 and a population of about 7,611,000 in 2013.
It has a temperate monsoon climate, with hot summers and mild winters. We retrieved the daily counts of BD among children aged 0–14 years, which accounted for more than 94.0 % of the total BD, during 1 January 2006 and 31 December 2012 from the Hefei Center for Disease Control and Prevention. The clinical diagnostic criteria for bacillary dysentery have been provided in a program published by the Chinese Center for Disease Control and Prevention in 2005 (China CDC 2005). According to the notifiable infectious disease regulations in China, information of bacillary dysentery cases, such as name, sex, age, occupation, telephone number, address, and date of onset, should be reported online to the Chinese Center for Disease Control and Prevention within 6 and 12 h of diagnosis using a unified format in towns and countryside, respectively. The cases were diagnosed by a professional clinician. What we received was an Excel document after de-identifying, and all the information was complete. Two of us were responsible for sorting data and extracting what we needed, such as sex, age, date of onset, and the International Classification of Disease, 10th version. Another one checked it until we reached consensus on the final data. The cases caused by bacteria were coded as bacillary dysentery (ICD-10 codes: A03) among children aged 0– 14 years were included in the data analysis. Daily meteorological data, including mean, maximum, and minimum temperature and relativity humidity in Hefei covering the same period were retrieved from the documentation of the Hefei Meteorological Bureau. DTR was computed by subtracting the daily minimum temperature from the daily maximum temperature. Statistical analyses Late spring and early autumn (May–October) were chosen as the main study period because it has the highest incidence of childhood BD. Lots of previous researches have demonstrated that the effect of DTR on human health is delayed (Chu et al. 2011; Song et al. 2008). Moreover, the relationship between DTR and diarrhea has been shown to be non-linear (Xu et al. 2013b). As a consequence, a distributed lag non-linear model was applied to incorporate both the lagged and nonlinear DTR effects (Gasparrini et al. 2010). A Poisson generalized linear regression model combined with a distributed lag non-linear model was used to quantify the effect of DTR on childhood BD. In this model, average temperature and relative humidity were controlled by using a natural cubic spline with 3 degrees of freedom (df). A natural cubic spline with 4 df per year was incorporated to control for long-term trends and seasonal patterns. We adjusted for the day of the
Int J Biometeorol
week as dummy variables in the model. In all cases, to evaluate the choice of df and model fit, the Akaike
Information Criterion (AIC) and an analysis of residuals were applied.
Y t e Poisson ðμt Þ Logðμt Þ ¼ a þ bDTRt;l þ nsðMean T t ; 3Þ þ ns RHt; 3 þ ns Timet ; 4 þ cDOWt
Where t is the day of the observation, Yt is the observed daily childhood bacillary dysentery on day t, and a is the model intercept. DTRt,l is a matrix obtained by applying the distributed lag non-linear model (DLNM) to DTR, b is vector of coefficients for DTRt,l, and l is the lag days. ns denotes the natural cubic spline. c indicates the vector of coefficient for day of the week (DOW)t. We found the effect of DTR on childhood BD was negligible for lags above 5 days. Furthermore, we examined the cumulative effects of DTR at lag 0–4 days using a lagstratified approach. The reason why we particularly concentrated on this period is that the AIC value of this model is relatively lower. Effects of DTR on childhood bacillary dysentery were estimated and presented as relative risks, stated for a 5 °C increase in DTR above 8 °C, which was selected by taking the average DTR level covering the whole study period. The R statistical environment (version 3.1.0) with the Bdlnm^ package (Gasparrini and Armstrong 2011) was used to run the overall data analysis. To compare the results, we performed a sensitivity analysis by altering the df for time, temperature and relative humidity.
Results There were 5544 cases of BD among children aged 0–14 years in Hefei across the whole study period. Table 1 shows the Table 1
summary statistics for meteorological variables and total and age- and gender-specific childhood bacillary dysentery cases. The average number of daily childhood BD was 4.3 (range 0– 15). The average values of daily mean temperature, relative humidity, and DTR were 24.6 °C (10.6–34.0 °C), 76.2 % (30.0–100.0 %), and 8.0 °C (1.2–18.1 °C). There were 658 days, with each day’s DTR over 8 °C during the study period. Figure 1a presents the temporal distribution of the number of childhood BD from 1 January 2006 to 31 December 2012 in Hefei, China, showing a strong seasonal trend. Figure 1b indicates that the number of childhood BD was greater in the months of May to October than other months. Figure 2 shows the time-series distribution of daily mean temperature, relative humidity, and diurnal temperature range from 1 January 2006 to 31 December 2012 in Hefei, showing an apparent seasonality in these variables. Figure 3 shows the cumulative effects of DTR on the total and gender-specific childhood BD. It presents that DTR was evidently associated with childhood BD. However, it was more likely to affect male children. Furthermore, the effect of DTR increased significantly above a DTR of 8 °C. Figure 4 illustrates the cumulative effects of DTR on agespecific childhood BD. It suggests that DTR was more likely to affect BD among 0–5-year-old children, with the relative risk increasing rapidly when DTR was over 8 °C. Table 2 depicts the spearman correlations among the weather variables. DTR was positively correlated with mean
Summary statistic for daily climate variables and daily childhood bacillary dysentery in Hefei, China, from 2006 to 2012
Variables
Mean temperature (°C) Maximum temperature (°C) Minimum temperature (°C) Diurnal temperature range (°C) Relative humidity (%) Cases of childhood bacillary dysentery 0–5 bacillary dysentery 6–14 bacillary dysentery Male bacillary dysentery Female bacillary dysentery
Mean
24.6 29.0 21.0 8.0 76.2 4.3 3.7 0.6 2.8 1.5
SD
4.4 4.6 4.6 2.9 12.1 2.7 2.4 1.0 2.0 1.4
Range
10.6–34.0 13.5–38.8 4.6–30.6 1.2–18.1 30.0–100.0 0.0–15.0 0.0–13.0 0.0–8.0 0.0–12.0 0.0–8.0
Percentile 25
50
75
21.6 26.0 17.7 6.0 69.0 2.0 2.0 0.0 1.0 0.0
24.9 29.2 21.6 8.0 77.0 4.0 3.0 0.0 2.0 1.0
28.1 32.7 24.8 9.8 85.0 6.0 5.0 1.0 4.0 2.0
Int J Biometeorol
6) for temperature and relative humidity. The results changed little (results not shown).
Discussion
Fig. 1 The daily and monthly distribution of childhood bacillary dysentery cases in Hefei, China, during 2006–2012
temperature (r=0.05), but negatively associated with humidity (r=−0.61). Table 3 represents the association between DTR and bacillary dysentery of total and age- and gender-specific children with and without controlling for mean temperature in two models in Hefei, China, during 2006–2012. In the model controlling for mean temperature, the effect was greatest at 1-day lag and a 5 °C increase in DTR was associated with an 8.0 % (95 % confidence interval (CI)=2.9–13.4 %) increase of bacillary dysentery cases for children aged 0–14 years. Nevertheless, the effect was greatest at 2-day lag and there was an 8.3 % (95 % CI=1.7–15.2 %) and 10.8 % (95 % CI=3.2– 19.0 %) increase in 0–5 years and male children cases, respectively, along with 5 °C increase in DTR. Not likewise, in the model without controlling for mean temperature, the effect was greatest at 2-day lag and a 5 °C increase in DTR was associated with an 8.8 % (95 % CI=2.9–15.0 %) increase of cases for children age 0–14 years. The same as the first model, the effect was greatest at 2-day lag and there was a 9.6 % (3.3– 16.4 %) and 12.8 % (5.3–20.8 %) increase in 0–5 years and male children cases, respectively. Sensitivity analyses were conducted by varying the df (4–6 per year) for time to control for season and altering the df (3–
This study yielded several interesting findings. Firstly, we found that there was a statistically significant association between DTR and childhood BD in Hefei during the study period. Then, the DTR effect increased significantly when DTR was above 8 °C. Besides, it was greatest for 1-day lag for total children and 2-day lag for children aged 0–5 years and male children. What’s more, the current study indicated that male children and children aged 0–5 years appeared to be more sensitive to the DTR effect than other children. The results of this study support our hypothesis that there was a predominant association between DTR and childhood BD. Xu et al. found that within-day variations of temperature were associated with emergency department visits for childhood diarrhea (Xu et al. 2013b), a kind of intestinal infectious disease mainly caused by bacteria or virus, which is consistent with our results. For the present, the exact mechanism by which exposure to large DTR levels can increase BD risk in children still remains indistinct. Nevertheless, there seems to be some reasonable explanations for their relationship. Bull declared that changes in weather might influence either humoral or cellular immunity (Bull 1980). Additionally, younger children have a relatively less developed immune system (Gerba et al. 1996) and lower self-care ability (Xu et al. 2012), meaning that they can not protect themselves from unsafe environment and thus may give rise to a greater vulnerability to temperature variations. Prior studies have reported that sudden changes in the temperature of inhaled air are correlated with the release of inflammatory mediators by mast cells (Graudenz et al. 2006; Togias et al. 1985), which could also account for some of the mechanisms related to higher bacillary dysentery prevalence (Feng et al. 2007; Ramsay et al. 2010). In the present study, the DTR effect increased rapidly above 8 °C, highlighting that both caregivers and health practitioners should be made aware of the potential threat posed by large DTR to children. During the whole study period, children in Hefei were exposed to relatively large DTR (>8 °C) for about 658 days. In the matter of the lag structure of DTR effect, we discovered that the greatest effect was at 1-day lag, which is in accordance with a previous study quantifying the relationship between DTR and morbidity caused by diarrhea (Xu et al. 2013b). Surprisingly, the findings of the two studies in different temperature zones were exactly similar. Bacillary dysentery is mainly transmitted through polluted drinking water and contaminated food or through person-toperson contact. Climate variations may directly impact the replication and survival of the pathogens in the environment
Int J Biometeorol Fig. 2 The time-series distribution of daily temperature, relative humidity, and diurnal temperature range from 2006 to 2012 in Hefei, China
Fig. 3 The overall effects of diurnal temperature range on bacillary dysentery among children aged 0–14 years in Hefei, China, from 2006 to 2012
Fig. 4 The overall effects of diurnal temperature range on childhood bacillary in Hefei, China, during 2006–2012
(Li et al. 2013a, b). It may also influence population behavior, including dietary pattern, eating habits, hygiene behavior, and susceptibility to different pathogen strains, which may indirectly affect the transmission of BD (Guan et al. 2008; Li et al. 2013a, b; Zhang et al. 2007, 2008). When it comes to the vulnerable populations, there existed various perspectives. Our study found that male children and children aged 0–5 years were more likely to suffer from BD during days with large DTR. With respect to gender, previous studies have inconsistent findings, which indicate that both males and females are vulnerable to DTR effect (Kan et al. 2007; Xu et al. 2013b). But the result of another study supports that male children are more vulnerable to the DTR effect (Xu et al. 2013a), which is consistent with our outcome. The
Table 2 Spearman’s correlation coefficient among daily weather conditions in Hefei, China, from 2006 to 2012
DTR Relative humidity
Mean temperature
DTR
0.05* 0.13**
−0.61**
*p value
ORIGINAL PAPER
The association between diurnal temperature range and childhood bacillary dysentery Li-ying Wen 1 & Ke-fu Zhao 2 & Jian Cheng 1 & Xu Wang 1 & Hui-hui Yang 1 & Ke-sheng Li 1 & Zhi-wei Xu 3,4 & Hong Su 1
Received: 12 June 2014 / Revised: 22 May 2015 / Accepted: 26 May 2015 # ISB 2015
Abstract Previous studies have found that mean, maximum, and minimum temperatures were associated with bacillary dysentery (BD). However, little is known about whether the within-day variation of temperature has any impact on bacillary dysentery. The current study aimed to identify the relationship between diurnal temperature range (DTR) and BD in Hefei, China. Daily data on BD counts among children aged 0–14 years from 1 January 2006 to 31 December 2012 were retrieved from Hefei Center for Disease Control and Prevention. Daily data on ambient temperature and relative humidity covering the same period were collected from the Hefei Bureau of Meteorology. A Poisson generalized linear regression model combined with a distributed lag non-linear model (DLNM) was used in the analysis after controlling the effects of season, long-term trends, mean temperature, and relative humidity. The results showed that there existed a statistically Li-ying Wen and Ke-fu Zhao contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s00484-015-1023-9) contains supplementary material, which is available to authorized users. * Hong Su [email protected] 1
Department of Epidemiology and Health Statistics, School of Public Health, Anhui Medical University, 81 Meishan Road, Hefei 230032, Anhui, China
2
Hefei Center for Disease Control and Prevention of Anhui Province, Hefei 230061, Anhui, China
3
School of Public Health and Social Work, Queensland University of Technology, Kelvin Grove, Brisbane, QLD 4509, Australia
4
Institute of Health and Biomedical Innovation, Queensland University of Technology, Kelvin Grove, Brisbane, QLD 4509, Australia
significant relationship between DTR and childhood BD. The DTR effect on childhood bacillary dysentery increased when DTR was over 8 °C. And it was greatest at 1-day lag, with an 8 % (95 % CI=2.9–13.4 %) increase of BD cases per 5 °C increment of DTR. Male children and children aged 0–5 years appeared to be more vulnerable to the DTR effect. The data indicate that large DTR may increase the incidence of childhood BD. Caregivers and health practitioners should be made aware of the potential threat posed by large DTR. Therefore, DTR should be taken into consideration when making targeted health policies and programs to protect children from being harmed by climate impacts. Keywords Climate change . Diurnal temperature range . Childhood . Bacillary dysentry . Hefei
Introduction Bacillary dysentery (BD), a bacterial infection of the mucosal surface of the intestines caused by different species of Shigella bacteria, including Shigella dysenteriae, Shigella flexneri, Shigella boydii, and Shigella sonnei (Niyogi. 2005), often results in severe diarrhea. Patients who are infected with Shigella frequently develop bloody diarrhea, fever, abdominal cramps, and tenesmus with an incubation period of 1 or 2 days (Niyogi 2005). The transmission of BD usually occurs via polluted drinking water and contaminated food or through person-toperson contact. It is a public health issue in both developed and developing countries (Zhang et al. 2008), particularly in less developed countries (Kotloff et al. 1999; von Seidlein et al. 2006; Walker et al. 2012; Wang et al. 2005, 2006). Currently, no specific drug or targeted vaccine is available for BD (Phalipon et al. 2008). But many researchers have been trying to develop a vaccine even in the face of many
Int J Biometeorol
challenges and difficulties (Barry et al. 2013; Camacho et al. 2013; Kim et al. 2013). In particular, the Chinese government has made great efforts in the prevention and control of BD, and has made obvious progresses. According to the National Notifiable Disease of China, the reported BD cases decreased from 205,972 in 2012 to 187,303 in 2013 (Ministry of Health, P.R. China.). Although the incidence of BD has declined largely, a considerable disease burden still exists (Zhang et al. 2012), especially among children (Barry et al. 2013; Ghaemi et al. 2007; Wang et al. 2005, 2006). Thus, it can be seen that BD in children would attract increasing attention. A variety of determinants, including environmental, nutritional, socioeconomic, and behavioral factors, can promote the occurrence of BD (Boru et al. 2013; Haley et al. 2010). Actually, in recent years, the meteorological factors have caused more and more research attention (Li et al. 2013a, b; Luo et al. 2010) and mean, maximum, and minimum temperatures have been associated with the occurrence of BD (Gao et al. 2014; Guan et al. 2008; Li et al. 2013a, b; Zhang et al. 2007, 2008). Amazingly, another meteorological indicator, e.g., diurnal temperature range (DTR), refined as the difference between maximal and minimal temperatures within 1 day, has sparked extraordinary interests. The reason may be that DTR can incorporate the information of maximum and minimum temperatures and reflect whether the weather is stable or not (Kalnay and Cai 2003). Previous researches have identified that DTR is correlated with respiratory and cardiovascular diseases (Chen et al. 2007; Lim et al. 2012a, b; Song et al. 2008; Xu et al. 2013a). However, the influence of temperature variation on BD has received little attention so far. Additionally, few studies have investigated the effect of DTR on children’s health (Xu et al. 2013b), and little is known about the relationship between DTR and childhood BD. In the light of the above considerations, we hypothesized that DTR was correlated with the incidence of childhood BD. This study emphasized four key research issues: (1) What is the relationship between DTR and BD? (2) Which children subgroups are most prone to DTR effect? (3) Is there any difference of DTR effect on childhood BD between two genders? (4) Is there a delayed effect of DTR on childhood BD?
Materials and methods Data collection Hefei is the capital city of Anhui province, located on the east-central of China (31° 52 N, 117° 17 E). It is compromised of nine counties, with a total land area of about 11,408 km2 and a population of about 7,611,000 in 2013.
It has a temperate monsoon climate, with hot summers and mild winters. We retrieved the daily counts of BD among children aged 0–14 years, which accounted for more than 94.0 % of the total BD, during 1 January 2006 and 31 December 2012 from the Hefei Center for Disease Control and Prevention. The clinical diagnostic criteria for bacillary dysentery have been provided in a program published by the Chinese Center for Disease Control and Prevention in 2005 (China CDC 2005). According to the notifiable infectious disease regulations in China, information of bacillary dysentery cases, such as name, sex, age, occupation, telephone number, address, and date of onset, should be reported online to the Chinese Center for Disease Control and Prevention within 6 and 12 h of diagnosis using a unified format in towns and countryside, respectively. The cases were diagnosed by a professional clinician. What we received was an Excel document after de-identifying, and all the information was complete. Two of us were responsible for sorting data and extracting what we needed, such as sex, age, date of onset, and the International Classification of Disease, 10th version. Another one checked it until we reached consensus on the final data. The cases caused by bacteria were coded as bacillary dysentery (ICD-10 codes: A03) among children aged 0– 14 years were included in the data analysis. Daily meteorological data, including mean, maximum, and minimum temperature and relativity humidity in Hefei covering the same period were retrieved from the documentation of the Hefei Meteorological Bureau. DTR was computed by subtracting the daily minimum temperature from the daily maximum temperature. Statistical analyses Late spring and early autumn (May–October) were chosen as the main study period because it has the highest incidence of childhood BD. Lots of previous researches have demonstrated that the effect of DTR on human health is delayed (Chu et al. 2011; Song et al. 2008). Moreover, the relationship between DTR and diarrhea has been shown to be non-linear (Xu et al. 2013b). As a consequence, a distributed lag non-linear model was applied to incorporate both the lagged and nonlinear DTR effects (Gasparrini et al. 2010). A Poisson generalized linear regression model combined with a distributed lag non-linear model was used to quantify the effect of DTR on childhood BD. In this model, average temperature and relative humidity were controlled by using a natural cubic spline with 3 degrees of freedom (df). A natural cubic spline with 4 df per year was incorporated to control for long-term trends and seasonal patterns. We adjusted for the day of the
Int J Biometeorol
week as dummy variables in the model. In all cases, to evaluate the choice of df and model fit, the Akaike
Information Criterion (AIC) and an analysis of residuals were applied.
Y t e Poisson ðμt Þ Logðμt Þ ¼ a þ bDTRt;l þ nsðMean T t ; 3Þ þ ns RHt; 3 þ ns Timet ; 4 þ cDOWt
Where t is the day of the observation, Yt is the observed daily childhood bacillary dysentery on day t, and a is the model intercept. DTRt,l is a matrix obtained by applying the distributed lag non-linear model (DLNM) to DTR, b is vector of coefficients for DTRt,l, and l is the lag days. ns denotes the natural cubic spline. c indicates the vector of coefficient for day of the week (DOW)t. We found the effect of DTR on childhood BD was negligible for lags above 5 days. Furthermore, we examined the cumulative effects of DTR at lag 0–4 days using a lagstratified approach. The reason why we particularly concentrated on this period is that the AIC value of this model is relatively lower. Effects of DTR on childhood bacillary dysentery were estimated and presented as relative risks, stated for a 5 °C increase in DTR above 8 °C, which was selected by taking the average DTR level covering the whole study period. The R statistical environment (version 3.1.0) with the Bdlnm^ package (Gasparrini and Armstrong 2011) was used to run the overall data analysis. To compare the results, we performed a sensitivity analysis by altering the df for time, temperature and relative humidity.
Results There were 5544 cases of BD among children aged 0–14 years in Hefei across the whole study period. Table 1 shows the Table 1
summary statistics for meteorological variables and total and age- and gender-specific childhood bacillary dysentery cases. The average number of daily childhood BD was 4.3 (range 0– 15). The average values of daily mean temperature, relative humidity, and DTR were 24.6 °C (10.6–34.0 °C), 76.2 % (30.0–100.0 %), and 8.0 °C (1.2–18.1 °C). There were 658 days, with each day’s DTR over 8 °C during the study period. Figure 1a presents the temporal distribution of the number of childhood BD from 1 January 2006 to 31 December 2012 in Hefei, China, showing a strong seasonal trend. Figure 1b indicates that the number of childhood BD was greater in the months of May to October than other months. Figure 2 shows the time-series distribution of daily mean temperature, relative humidity, and diurnal temperature range from 1 January 2006 to 31 December 2012 in Hefei, showing an apparent seasonality in these variables. Figure 3 shows the cumulative effects of DTR on the total and gender-specific childhood BD. It presents that DTR was evidently associated with childhood BD. However, it was more likely to affect male children. Furthermore, the effect of DTR increased significantly above a DTR of 8 °C. Figure 4 illustrates the cumulative effects of DTR on agespecific childhood BD. It suggests that DTR was more likely to affect BD among 0–5-year-old children, with the relative risk increasing rapidly when DTR was over 8 °C. Table 2 depicts the spearman correlations among the weather variables. DTR was positively correlated with mean
Summary statistic for daily climate variables and daily childhood bacillary dysentery in Hefei, China, from 2006 to 2012
Variables
Mean temperature (°C) Maximum temperature (°C) Minimum temperature (°C) Diurnal temperature range (°C) Relative humidity (%) Cases of childhood bacillary dysentery 0–5 bacillary dysentery 6–14 bacillary dysentery Male bacillary dysentery Female bacillary dysentery
Mean
24.6 29.0 21.0 8.0 76.2 4.3 3.7 0.6 2.8 1.5
SD
4.4 4.6 4.6 2.9 12.1 2.7 2.4 1.0 2.0 1.4
Range
10.6–34.0 13.5–38.8 4.6–30.6 1.2–18.1 30.0–100.0 0.0–15.0 0.0–13.0 0.0–8.0 0.0–12.0 0.0–8.0
Percentile 25
50
75
21.6 26.0 17.7 6.0 69.0 2.0 2.0 0.0 1.0 0.0
24.9 29.2 21.6 8.0 77.0 4.0 3.0 0.0 2.0 1.0
28.1 32.7 24.8 9.8 85.0 6.0 5.0 1.0 4.0 2.0
Int J Biometeorol
6) for temperature and relative humidity. The results changed little (results not shown).
Discussion
Fig. 1 The daily and monthly distribution of childhood bacillary dysentery cases in Hefei, China, during 2006–2012
temperature (r=0.05), but negatively associated with humidity (r=−0.61). Table 3 represents the association between DTR and bacillary dysentery of total and age- and gender-specific children with and without controlling for mean temperature in two models in Hefei, China, during 2006–2012. In the model controlling for mean temperature, the effect was greatest at 1-day lag and a 5 °C increase in DTR was associated with an 8.0 % (95 % confidence interval (CI)=2.9–13.4 %) increase of bacillary dysentery cases for children aged 0–14 years. Nevertheless, the effect was greatest at 2-day lag and there was an 8.3 % (95 % CI=1.7–15.2 %) and 10.8 % (95 % CI=3.2– 19.0 %) increase in 0–5 years and male children cases, respectively, along with 5 °C increase in DTR. Not likewise, in the model without controlling for mean temperature, the effect was greatest at 2-day lag and a 5 °C increase in DTR was associated with an 8.8 % (95 % CI=2.9–15.0 %) increase of cases for children age 0–14 years. The same as the first model, the effect was greatest at 2-day lag and there was a 9.6 % (3.3– 16.4 %) and 12.8 % (5.3–20.8 %) increase in 0–5 years and male children cases, respectively. Sensitivity analyses were conducted by varying the df (4–6 per year) for time to control for season and altering the df (3–
This study yielded several interesting findings. Firstly, we found that there was a statistically significant association between DTR and childhood BD in Hefei during the study period. Then, the DTR effect increased significantly when DTR was above 8 °C. Besides, it was greatest for 1-day lag for total children and 2-day lag for children aged 0–5 years and male children. What’s more, the current study indicated that male children and children aged 0–5 years appeared to be more sensitive to the DTR effect than other children. The results of this study support our hypothesis that there was a predominant association between DTR and childhood BD. Xu et al. found that within-day variations of temperature were associated with emergency department visits for childhood diarrhea (Xu et al. 2013b), a kind of intestinal infectious disease mainly caused by bacteria or virus, which is consistent with our results. For the present, the exact mechanism by which exposure to large DTR levels can increase BD risk in children still remains indistinct. Nevertheless, there seems to be some reasonable explanations for their relationship. Bull declared that changes in weather might influence either humoral or cellular immunity (Bull 1980). Additionally, younger children have a relatively less developed immune system (Gerba et al. 1996) and lower self-care ability (Xu et al. 2012), meaning that they can not protect themselves from unsafe environment and thus may give rise to a greater vulnerability to temperature variations. Prior studies have reported that sudden changes in the temperature of inhaled air are correlated with the release of inflammatory mediators by mast cells (Graudenz et al. 2006; Togias et al. 1985), which could also account for some of the mechanisms related to higher bacillary dysentery prevalence (Feng et al. 2007; Ramsay et al. 2010). In the present study, the DTR effect increased rapidly above 8 °C, highlighting that both caregivers and health practitioners should be made aware of the potential threat posed by large DTR to children. During the whole study period, children in Hefei were exposed to relatively large DTR (>8 °C) for about 658 days. In the matter of the lag structure of DTR effect, we discovered that the greatest effect was at 1-day lag, which is in accordance with a previous study quantifying the relationship between DTR and morbidity caused by diarrhea (Xu et al. 2013b). Surprisingly, the findings of the two studies in different temperature zones were exactly similar. Bacillary dysentery is mainly transmitted through polluted drinking water and contaminated food or through person-toperson contact. Climate variations may directly impact the replication and survival of the pathogens in the environment
Int J Biometeorol Fig. 2 The time-series distribution of daily temperature, relative humidity, and diurnal temperature range from 2006 to 2012 in Hefei, China
Fig. 3 The overall effects of diurnal temperature range on bacillary dysentery among children aged 0–14 years in Hefei, China, from 2006 to 2012
Fig. 4 The overall effects of diurnal temperature range on childhood bacillary in Hefei, China, during 2006–2012
(Li et al. 2013a, b). It may also influence population behavior, including dietary pattern, eating habits, hygiene behavior, and susceptibility to different pathogen strains, which may indirectly affect the transmission of BD (Guan et al. 2008; Li et al. 2013a, b; Zhang et al. 2007, 2008). When it comes to the vulnerable populations, there existed various perspectives. Our study found that male children and children aged 0–5 years were more likely to suffer from BD during days with large DTR. With respect to gender, previous studies have inconsistent findings, which indicate that both males and females are vulnerable to DTR effect (Kan et al. 2007; Xu et al. 2013b). But the result of another study supports that male children are more vulnerable to the DTR effect (Xu et al. 2013a), which is consistent with our outcome. The
Table 2 Spearman’s correlation coefficient among daily weather conditions in Hefei, China, from 2006 to 2012
DTR Relative humidity
Mean temperature
DTR
0.05* 0.13**
−0.61**
*p value
