• View in gallery

    Number of hemorrhagic fever with renal syndrome cases and climate variables in Heilongjiang Province, China, 2001–2009.

  • View in gallery

    A, Autocorrelation and B, partial autocorrelation of residuals of the seasonal autoregressive integrated moving average model (SARIMA) model, Heilongjiang Province, China, 2001–2009.

  • View in gallery

    Forecasted and observed number of hemorrhagic fever with renal syndrome (HFRS) cases based on a seasonal autoregressive integrated moving average model of climatic variation in Heilongjiang Province, China, 2001–2009.

  • 1.

    Ministry of Health, 1998. Handbook of Epidemic Hemorrhagic Fever Prevention and Control. Beijing: China People's Health Publishing House, 6380.

    • Search Google Scholar
    • Export Citation
  • 2.

    Schmaljohn C, Hjelle B, 1997. Hantaviruses: a global disease problem. Emerg Infect Dis 3: 95104.

  • 3.

    Bi Z, Formenty PB, Roth CE, 2008. Hantavirus infection: a review and global update. J Infect Dev Ctries 2: 323.

  • 4.

    Yan L, Fang LQ, Huang HG, Zhang LQ, Feng D, Zhao WJ, Zhang WY, Li XW, Cao WC, 2007. Landscape elements and Hantaan virus-related hemorrhagic fever with renal syndrome, People's Republic of China. Emerg Infect Dis 13: 13011306.

    • Search Google Scholar
    • Export Citation
  • 5.

    Jiang JF, Wu XM, Zuo SQ, Wang RM, Chen LQ, Wang BC, Dun Z, Zhang PH, Guo TY, Cao WC, 2006. Study on the association between Hantavirus infection and Rattus norvegicus. Chin J Epidemiol 27: 196199.

    • Search Google Scholar
    • Export Citation
  • 6.

    Zhang WY, Fang LQ, Jiang JF, Hui FM, Glass GE, Yan L, Xu YF, Zhao WJ, Yang H, Liu W, Cao WC, 2009. Predicting the risk of Hantavirus infection in Beijing, People's Republic of China. Am J Trop Med Hyg 80: 678683.

    • Search Google Scholar
    • Export Citation
  • 7.

    Bi P, Wu X, Zhang F, Parton K, Tong S, 1998. Seasonal rainfall variability, the incidence of hemorrhagic fever with renal syndrome, and prediction of the disease in low-lying areas of China. Am J Epidemiol 148: 276281.

    • Search Google Scholar
    • Export Citation
  • 8.

    Bi P, Tong S, Donald K, Parton K, Ni J, 2002. Climatic, reservoir and occupational variables and the transmission of haemorrhagic fever with renal syndrome in China. Int J Epidemiol 31: 189193.

    • Search Google Scholar
    • Export Citation
  • 9.

    Ernest SKM, Brown JH, Parmenter RR, 2000. Rodents, plants, and precipitation: spatial and temporal dynamics of consumers and resources. Oikos 88: 470482.

    • Search Google Scholar
    • Export Citation
  • 10.

    Glass GE, Shields T, Cai B, Yates TL, Parmenter R, 2007. Persistently highest risk areas for hantavirus pulmonary syndrome: potential sites for refugia. Ecol Appl 17: 129139.

    • Search Google Scholar
    • Export Citation
  • 11.

    Langlois JP, Fahrig L, Merriam G, Artsob H, 2001. Landscape structure infuences continental distribution of Hantavirus in deer mice. Landscape Ecol 16: 255266.

    • Search Google Scholar
    • Export Citation
  • 12.

    Madsen T, Shine R, 1999. Rainfall and rats: climatically-driven dynamics of a tropical rodent population. Aust J Ecol 24: 8089.

  • 13.

    Engelthaler DM, Mosley DG, Cheek JE, Levy CE, Komatsu KK, Ettestad P, Davis T, Tanda DT, Miller L, Frampton JW, Porter R, Bryan RT, 1999. Climatic and environmental patterns associated with hantavirus pulmonary syndrome, Four Corners region, United States. Emerg Infect Dis 5: 8794.

    • Search Google Scholar
    • Export Citation
  • 14.

    Luis AD, Douglass RJ, Mills JN, Bjørnstad ON, 2010. The effect of seasonality, density and climate on the population dynamics of Montana deer mice, important reservoir hosts for Sin Nombre Hantavirus. J Anim Ecol 79: 462470.

    • Search Google Scholar
    • Export Citation
  • 15.

    Glass GE, Cheek JE, Patz JA, Shields TM, Doyle TJ, Thoroughman DA, Hunt DK, Enscore RE, Gage KL, Irland C, Peters CJ, Bryan R, 2000. Using remotely sensed data to identify areas at risk for Hantavirus pulmonary syndrome. Emerg Infect Dis 6: 238247.

    • Search Google Scholar
    • Export Citation
  • 16.

    Hjelle B, Glass GE, 2000. Outbreak of Hantavirus infection in the Four Corners region of the United States in the wake of the 1997–1998 El Niño-Southern Oscillation. J Infect Dis 181: 15691573.

    • Search Google Scholar
    • Export Citation
  • 17.

    Tamerius JD, Wise EK, Uejio CK, McCoy AL, Comrie AC, 2007. Climate and human health: synthesizing environmental complexity and uncertainty. Stochastic Environ Res Risk Assess 21: 601613.

    • Search Google Scholar
    • Export Citation
  • 18.

    Tersago K, Verhagen R, Servais A, Heyman P, Ducoffre G, Leirs H, 2009. Hantavirus disease (nephropathia epidemica) in Belgium: effects of tree seed production and climate. Epidemiol Infect 137: 250256.

    • Search Google Scholar
    • Export Citation
  • 19.

    Klempa B, 2009. Hantaviruses and climate change. Clin Microbiol Infect 15: 518523.

  • 20.

    Bi P, Parton K, 2003. El Niño and incidence of hemorrhagic fever with renal syndrome in China. JAMA 289: 176177.

  • 21.

    Fang LQ, Wang XJ, Liang S, Yan LL, Song SX, Zhang WY, Qian Q, Li YP, Wei L, Wang ZQ, Yang H, Cao WC, 2010. Spatiotemporal trends and climatic factors of hemorrhagic fever with renal syndrome epidemic in Shandong Province, China. PLoS Negl Trop Dis 4: 110.

    • Search Google Scholar
    • Export Citation
  • 22.

    Pettersson L, Boman J, Juto P, Evander M, Ahlm C, 2008. Outbreak of Puumala virus infection, Sweden. Emerg Infect Dis 14: 808810.

  • 23.

    Schwarz AC, Ranft U, Piechotowski I, Childs JE, Brockmann SO, 2009. Risk factors for human infection with Puumala virus, southwestern Germany. Emerg Infect Dis 15: 10321039.

    • Search Google Scholar
    • Export Citation
  • 24.

    Huang RH, Wu YF, 1989. The infuence of ENSO on the summer climate change in China and its mechanism. Adv Atmos Sci 6: 2132.

  • 25.

    Zhang WY, Guo WD, Fang LQ, Li CP, Bi P, Glass GE, Jiang JF, Sun SH, Qian Q, Liu W, Yan L, Yang H, Tong SL, Cao WC, 2010. Climate variability and hemorrhagic fever with renal syndrome transmission in northeastern China. Environ Health Perspect 118: 915920.

    • Search Google Scholar
    • Export Citation
  • 26.

    National Centers for Environmental Prediction, 2010. Climatic Forecast System. Available at: http://www.cpc.ncep.noaa.gov/data/indices/soi. Accessed June 6, 2010.

    • Search Google Scholar
    • Export Citation
  • 27.

    Hu W, Tong S, Mengersen K, Connell D, 2007. Weather variability and the incidence of cryptosporidiosis: comparison of time series poisson regression and SARIMA models. Ann Epidemiol 17: 679688.

    • Search Google Scholar
    • Export Citation
  • 28.

    Allard R, 1998. Use of time-series analysis in infectious disease surveillance. Bull World Health Organ 76: 327333.

  • 29.

    Box G, Jenkins G, 1970. Time-Series Analysis: Forecasting and Control. San Francisco, CA: Holden-Day, 376382.

  • 30.

    Hu W, Mengersen K, Bi P, Tong S, 2007. Time series analysis of the risk factors for hemorrhagic fever with renal syndrome: comparison of statistical models. Epidemiol Infect 135: 245252.

    • Search Google Scholar
    • Export Citation
  • 31.

    Bozdogan H, 1987. Model-selection and Akaike's information criterion (AIC): the general theory and its analytical extensions. Psychometrika 52: 345370.

    • Search Google Scholar
    • Export Citation
  • 32.

    Liddle AR, 2007. Information criteria for astrophysical model selection. Mon Not R Astron Soc 377: 7478.

  • 33.

    SAS, 2008. SAS/ETS®9.2: User's Guide. Cary, NC: SAS Institute Inc., 177300.

  • 34.

    Calisher CH, Wagoner KD, Amman BR, Root JJ, Douglass RJ, Kuenzi AJ, Abbott KD, Parmenter C, Yates TL, Ksiazek TG, Beaty BJ, Mills JN, 2007. Demographic factors associated with prevalence of antibody to Sin Nombre virus in deer mice in the western United States. J Wildl Dis 43: 111.

    • Search Google Scholar
    • Export Citation
  • 35.

    Glass GE, Livingstone W, Mills JN, Hlady WG, Fine JB, Biggler W, Coke T, Frazier D, Atherley S, Rollin PE, Ksiazek TG, Peters CJ, Childs JE, 1998. Black Creek Canal virus infection in Sigmodon hispidus in southern Florida. Am J Trop Med Hyg 59: 699703.

    • Search Google Scholar
    • Export Citation
  • 36.

    Yahnke CJ, Meserve PL, Ksiazek TG, Mills JN. Patterns of infection with Laguna Negra virus in wild populations of Calomys laucha in the central Paraguayan Chaco 2001. Am J Trop Med Hyg 65: 768776.

    • Search Google Scholar
    • Export Citation
  • 37.

    Luo CW, Liu QY, Hou JL, 2009. Correlation analysis and regression model of epidemic factors of hemorrhagic fever with renal syndrome in Heihe City, Heilongjiang Province. Dis Surveill 24: 118120.

    • Search Google Scholar
    • Export Citation
  • 38.

    Thomson MC, Mason SJ, Phindela T, Connor SJ, 2005. Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana. Am J Trop Med Hyg 73: 214221.

    • Search Google Scholar
    • Export Citation
  • 39.

    Zhou G, Minakawa N, Githeko AK, Yan G, 2004. Association between climate variability and malaria epidemics in the east African highlands. Proc Natl Acad Sci USA 101: 23752380.

    • Search Google Scholar
    • Export Citation
  • 40.

    Cazelles B, Chavez M, McMichael AJ, Hales S, 2005. Nonstationary influence of El Niño on the synchronous dengue epidemics in Thailand. PLoS Med 2: e106.

    • Search Google Scholar
    • Export Citation
  • 41.

    Hales S, de Wet N, Maindonald J, Woodward A, 2002. Potential effect of population and climate changes on global distribution of dengue fever: an empirical model. Lancet 360: 830834.

    • Search Google Scholar
    • Export Citation
  • 42.

    Tong S, Hu W, 2001. Climate variation and incidence of Ross river virus in Cairns, Australia: a time-series analysis. Environ Health Perspect 109: 12711273.

    • Search Google Scholar
    • Export Citation
  • 43.

    Tong S, Hu W, 2002. Different responses of Ross River virus to climate variability between coastline and inland cities in Queensland, Australia. Occup Environ Med 59: 739744.

    • Search Google Scholar
    • Export Citation
  • 44.

    Chaves LF, Pascual M, 2006. Climate cycles and forecasts of cutaneous leishmaniasis, a nonstationary vector-borne disease. PLoS Med 3: 295.

  • 45.

    Nakazawa Y, Williams R, Peterson AT, Mead P, Staples E, Gage KL, 2007. Climate change effects on plague and tularemia in the United States. Vector Borne Zoonotic Dis 7: 529540.

    • Search Google Scholar
    • Export Citation
  • 46.

    Hallett TB, Coulson T, Pilkington JG, Clutton-Brock TH, Pemberton JM, Grenfell BT, 2004. Why large-scale climate indices seem to predict ecological processes better than local weather. Nature 430: 7175.

    • Search Google Scholar
    • Export Citation
  • 47.

    Zheng ZM, Jiang ZK, Chen AG, 2008. Rodents Zoology. Shanghai: Shanghai Jiaotong University Press, 400450.

 

 

 

 

 

Association between Hemorrhagic Fever with Renal Syndrome Epidemic and Climate Factors in Heilongjiang Province, China

View More View Less
  • Department of Health Statistics, College of Public Health, Tianjin Medical University, Tianjin, China; Institute of Disease Control and Prevention of PLA, Beijing, China; School of Population Health, Infectious Disease Epidemiology Unit, University of Queensland, Brisbane, Queensland, Australia; Tianjin Medical University General Hospital, Tianjin, China; Nankai University, Tianjin, China

The purpose of this study was to quantify the relationship between climate variation and transmission of hemorrhagic fever with renal syndrome (HFRS) in Heilongjiang Province, a highly endemic area for HFRS in China. Monthly notified HFRS cases and climatic data for 2001–2009 in Heilongjiang Province were collected. Using a seasonal autoregressive integrated moving average model, we found that relative humidity with a one-month lag (β = −0.010, P = 0.003) and a three-month lag (β = 0.008, P = 0.003), maximum temperature with a two-month lag (β = 0.082, P = 0.028), and southern oscillation index with a two-month lag (β = −0.048, P = 0.019) were significantly associated with HFRS transmission. Our study also showed that predicted values expected under the seasonal autoregressive integrated moving average model were highly consistent with observed values (Adjusted R2 = 83%, root mean squared error = 108). Thus, findings may help add to the knowledge gap of the role of climate factors in HFRS transmission in China and also assist national local health authorities in the development/refinement of a better strategy to prevent HFRS transmission.

Introduction

Hemorrhagic fever with renal syndrome (HFRS), is caused by different species of hantavirus transmitted by rodents, which can lead to severe clinical symptoms of fever, hemorrhage, and acute renal dysfunction.1 This disease is found in more than 30 countries and has become a major public health problem in China,2,3 It is endemic to 28 of 31 provinces in China, accounts for 90% of the HFRS cases reported globally, and has a case-fatality rate of 3–10%.4 In China, HFRS is caused mainly by two types of hantaviruses, Hantaan virus and Seoul virus, which coevolved with distinct rodent hosts. Hantaan virus is associated with Apodemus agrarius, whereas Seoul virus is associated with Rattus norvegicus and often causes a less severe symptome.1,4 Although prevention and control measures such as rodent control, vaccination, and environmental management have been implemented in parts of China most at risk, HFRS remains as a serious public health problem and there are approximately 20,000–50,000 new cases annually.

The incidence of HFRS in China varies geographically, seasonally, and interannually and is influenced by density of the rodent population, prevalence of hantavirus infection in rodents, and contact rate between rodents and humans.5,6 Although the infection rate and population dynamics of rodents are believed to be influenced by climatic factors,713 the association between climate and host populations is complex in most places.14

Studies in different areas have suggested that climate factors, such as temperature, precipitation, and relative humidity, may influence the incidence of HFRS.68 However, the role of climate factors in the transmission of HFRS differs in different regions.4,612 For example, increased grass seed production followed by heavy precipitation, as a result of the El Niño Southern Oscillation (ENSO), was found to be associated with higher Peromyscus maniculatus rodent density in southwestern United States and provides reservoirs for Sin Nombre virus.13,1517 However, this result was not duplicated in other areas in the United States.14 In addition, a study in west-central Europe indicated that increased mean temperatures had a positive relationship with the outbreak of Puumala hantavirus infections through high seed production and high bank vole densities. However, other studies in Scandinavia have shown that warm winters led to a decrease in vole populations as a result of the decrease in protective snow cover.18,19 Moreover, studies in eastern China showed that excessive precipitation could have a negative impact on rodents by destroying their habitats.8,20 In addition, temperature, vegetation type, land use, and ENSO have been shown to be associated with the transmission of HFRS in China.4,6,7,21 Therefore, climate variables related to rodent population dynamics may serve as indicators for the risk of HFRS transmission in humans.2224

Heilongjiang Province, which is a cold area in northeastern China, was the first identified epidemic foci of HFRS in China.2 In the past few years, the incidence of HFRS in Heilongjiang Province remains high. In 2009, the number of HFRS report cases reached 1,701, which accounted for 20% of all cases in China. Until now, there have been no studies conducted that quantified the relationship between climate variation and transmission of HFRS in Heilongjiang Province and few studies in China.7,8,20,21,25

The purpose of this study was to quantify the relationship between incidence of HFRS and climate in Heilongjiang Province. An additional objective was to develop a climate-based epidemic forecasting model to assist monitoring of HFRS incidence in the region to inform public health policy in Heilongjiang Province on the basis of changes in particular environmental variables.

Materials and Methods

Study areas.

Heilongjiang Province (121°11′–135°5′E, 43°25′–53°33′N) includes 13 cities and 66 counties with a population of 38.13 million. It has an area of 4546 km2, of which 51% is forest and 30% is farmland. The climate in Heilongjiang Province is a monsoon climate, with an average annual temperature between −4°C and 5°C. The annual temperature during the summer is 16–23°C, and rainfall during the summer accounts for 65% of all the total precipitation in a year. No significant difference in rainfall was found between southern and northern regions; this difference was greater between eastern and west regions. The annual average relative humidity is 60–70%, and there is a similar spatial distribution feature for relative humidity as for rainfall. In addition, there are 32 meteorologic stations in Heilongjiang Province.

Data collection.

Monthly numbers of HFRS cases in the 13 cities and 66 counties in Heilongjiang Province during January 2001–December 2009 were collected by the Center for Disease Control and Prevention of Heilongjiang Province. In this study, all the HFRS cases were confirmed according to the diagnosis from the Ministry of Health of the People's Republic of China.1 A confirmed case of HFRS was defined as illness in a person who had traveled to an HFRS-endemic area or who had come into contact with rodent feces, saliva, and urine within two months before onset of illness, and who had an acute illness characterized by abrupt onset of at least two of the following clinical features: fever, chills, hemorrhage, headache, back pain, abdominal pain, acute renal dysfunction, and hypotension. In addition, the person had to have had at least one of the laboratory criteria for diagnosis: a positive result for hantavirus-specific IgM, or a four-fold increase in titers of hantavirus-specific IgG, or a positive result for hantavirus-specific RNA acid by reverse transcription–polymerase chain reaction in clinical specimens, or hantavirus isolated from clinical specimens.1 Demographic data were obtained from the Heilongjiang Provincial Bureau of Statistics.

Monthly climatic data from 2001 through 2009 in Heilongjiang Province, including monthly mean temperature (°C), monthly maximum temperature (°C), monthly minimum temperature (°C), monthly relative humidity (%) and monthly rainfall (rainfall, mm), were obtained from the Chinese Bureau of Meteorology. Temperature was measured four times per day at 2:00 am, 8:00 am, noon, and 8:00 pm. Mean daily temperature is defined as the average of the four values. Mean monthly temperature is defined as the average of the mean daily temperature of every month. Maximum and minimum monthly and daily temperatures were also obtained. Monthly rainfall is defined as total precipitation every month, including rain and snow. Daily relative humidity is defined as average relative humidity at 2:00 am, 8:00 am, noon, and 8:00 pm, and monthly relative humidity is the average of the daily relative humidity of every month. In this study, we used the average of each climatic variable measured at the 32 meteorologic stations to explore the association between climate variables and incidence of HFRS by using the seasonal autoregressive integrated moving average (SARIMA) models. Southern oscillation index (SOI) was used as an indicator of ENSO, the most important coupled ocean–atmosphere phenomenon that affects the climate in China.24 The SOI was provided by the Climate Prediction Center of the National Weather Service.26

Statistical analysis.

Time series analysis has been extensively used to study the association between climate variability and disease transmission.25,27,28 More specifically, the SARIMA models are an available approach for interpreting and applying surveillance data in disease control, prevention, and forecast.29 The SARIMA models are especially useful in modeling the temporal-dependent structure of time series data with the adjustment of the impact of seasonal and interannual tendency.30

The SARIMA models have been successfully applied to predict the incidence of infectious diseases. A SARIMA model with nonseasonal and seasonal factors is denoted by ARIMA (p, d, q) × (P, D, Q)s, where the term (p, d, q) gives the order of nonseasonal part, (P, D, Q)s gives the order of seasonal part, and s is the number of observations in a seasonal cycle. The modeling of SARIMA involved identification, estimation, diagnostic checking, and forecasting stages, corresponding to the stages described by Box and Jenkins.29 The autocorrelation function (ACF) and partial autocorrelation function (PACF) were calculated to assess the appropriate orders of autocorrelation and moving average terms. Significance tests for parameter estimates indicate whether some terms in the model may be unnecessary. Goodness-of-fit statistics, such as R2, Akaike's Information Criterion (AIC), Schwarz Bayesian Information Criterion, root mean square error, mean absolute percent error and adjusted R2, aid in comparing this model to others. Tests for white noise residuals indicate whether the residual series contains additional information that might be used by a more complex model.

In this study, the climatic variables, including mean monthly temperature, maximum monthly temperature, minimum monthly temperature, monthly relative humidity, monthly rainfall, SOI, and corresponding lags of 1–6 months, were the final pool of candidate predictor variables. The best subsets procedure was applied to select predictors from candidate variables in this study. The general idea behind the best subsets method is that we select the subset of predictors that are best at meeting well-defined objective criterion, such as having the smallest AIC.31,32 In our study, all possible combinations of lags of climatic variables from one to six months were included to build different models, and the model with the smallest AIC value was considered the final model. The maximum-likelihood method was used for parameter estimation. The AIC was applied to compare goodness-of-fit within models. Furthermore, tests for white noise residuals with Box-Ljung test indicate whether the residual series contains additional information that might be used in a more complex model.

The number of HFRS cases was divided into two parts: data for January 2001–December 2008 were used to build the SARIMA model, and data for January–December 2009 were used to validate the model. A natural log transformation of the dependent variable, the monthly cases of HFRS, was conducted to meet the requirements of stationary and normality for the SARIMA model. In addition, the cubic root transformation was also applied to the monthly maximum temperature and SOI.25,27 In this study, data analysis was performed by using SAS version 9.2 software.33

Results

Descriptive analysis.

There were 30,439 HFRS cases during 2001–2009 and an annual average incidence of 8.87 cases/100,000 population. There was seasonality in the number of cases during the study period. One peak of HFRS cases in winter (October–December) was observed and accounted for 53.8% of cases. Another peak appeared in summer (May–July) and accounted for 19.8% of cases. During January 2001–December 2009, monthly mean temperature, monthly maximum temperature, monthly minimum temperature, monthly relative humidity, monthly rainfall, and monthly SOI ranged between −24.2°C and 22.5°C, −18.4°C and 27.9°C, −29.4°C and 17.9°C, 39% and 81.6%, 0.23 mm and 192.08 mm, and −4.1 and 2.7 respectively. The temporal variation of climatic factors and the number of cases during the study period is shown in Figure 1.

Figure 1.
Figure 1.

Number of hemorrhagic fever with renal syndrome cases and climate variables in Heilongjiang Province, China, 2001–2009.

Citation: The American Society of Tropical Medicine and Hygiene 89, 5; 10.4269/ajtmh.12-0473

Correlation analysis.

The bivariate correlation analysis between climatic variables and HFRS was conducted by using cross-correlation analysis adjusted for seasonality. There were significant lag effects between the climatic factors and the monthly cases of HFRS (Table 1). The cross-correlation coefficient reached the maximum at the three-month lag or four-month lag.

Table 1

Cross-correlation coefficients between climate variables and number of HFRS in Heilongjiang Province, China*

Climate variableLag 0Lag 1Lag 2Lag 3Lag 4Lag 5Lag 6
Mean T (°C)−0.36−0.060.200.370.430.410.32
Maximum T (°C)−0.37−0.060.210.360.420.410.32
Minimum T (°C)−0.34−0.060.200.380.450.420.31
Relative humidity (%)0.05−0.190.000.400.510.17−0.30
Mean rainfall (mm)−0.28−0.26−0.080.330.620.440.14
SOI−0.04−0.09−0.08−0.22−0.20−0.19−0.20

HFRS = hemorrhagic fever with renal syndrome; T = temperature; SOI = southern oscillation index.

P < 0.05.

Maximum cross-correlation coefficient in each row.

SARIMA model.

The results showed that first-order autoregression (β = 0.748, P = 0.0001) and seasonal autoregression (β = 0.887, P = 0.0001, lag = 12), relative humidity at one-month and three-month lags, monthly maximum temperature (curt root transformation) at two-month lag, and SOI (curt root transformation) at the two-month lag were significantly associated with transmission of HFRS, and other climatic variables, including monthly minimum temperature, rainfall, and monthly mean temperature, were not significantly associated (Table 2).

Table 2

SARIMA regression of number of HFRS in Heilongjiang Province, China*

ModelβSEtP95% CI
Constant5.2900.50210.5320.00014.306–6.274
Autoregression, lag 10.7480.07010.6240.00010.611–0.885
Seasonal autoregression, lag 120.8870.03724.0870.00010.814–0.960
Relative humidity (%), 1-month lag−0.0100.003−3.2660.002−0.016 to −0.004
Relative humidity (%), 3-month lag0.0080.0032.4260.0170.002–0.014
Curt root (MaxT), 2-month lag0.0820.0282.9170.0040.027–0.137
Curt root (SOI), 2-month lag−0.0480.019−2.6020.011−0.085 to −0.011

SARIMA = seasonal autoregressive integrated moving average model; HFRS = hemorrhagic fever with renal syndrome; CI = confidence interval; MaxT = monthly maximum temperature; SOI: southern oscillation index.

An autocorrelation check of residual with Box-Ljung test also showed that the model was appropriate (χ2 = 4.94, P = 0.2930). The ACF and PACF also showed that there was no significant autocorrelation between residuals at different lags in the final ARIMA model (Figure 2). The final model can be expressed as
DE1
Figure 2.
Figure 2.

A, Autocorrelation and B, partial autocorrelation of residuals of the seasonal autoregressive integrated moving average model (SARIMA) model, Heilongjiang Province, China, 2001–2009.

Citation: The American Society of Tropical Medicine and Hygiene 89, 5; 10.4269/ajtmh.12-0473

The number of monthly observed cases and that of fitted HFRS cases from the model in Heilongjiang Province was compared, and they matched reasonably well (Figure 3). The validation for January 2009 through December 2009 also showed good fitness between the observed and predicted values (Adjusted R2 = 83%, root mean square error = 108).

Figure 3.
Figure 3.

Forecasted and observed number of hemorrhagic fever with renal syndrome (HFRS) cases based on a seasonal autoregressive integrated moving average model of climatic variation in Heilongjiang Province, China, 2001–2009.

Citation: The American Society of Tropical Medicine and Hygiene 89, 5; 10.4269/ajtmh.12-0473

Discussion

This study demonstrated that the number of monthly notified HFRS cases was associated with climate variability occurring in the previous months. More importantly, after controlling the seasonal and autoregressive effect, our results showed an important seasonal signal in monthly maximum temperature, relative humidity, and SOI with a lag of 1–3 months in the association with reported HFRS cases. Identifying lag effects of climate variables included in our model will help local public health authorities to characterize the HFRS cases as they occur so as to reasonably forecast and prevent transmission of HFRS during epidemic seasons.

There is now a broad consensus that temperature factors are important in HFRS epidemics, but their association may vary among and within different areas.6,7,11,12 Some studies found that temperature had a positive effect,6,11 and others found a negative effect on HFRS incidence.7,12,13 The present study showed that monthly maximum temperature was positively associated with the monthly number of HFRS cases in this province, and other temperature factors, such as monthly minimum temperature and monthly mean temperature, were not included in the final model. A possible underlying mechanism for this finding is that temperature may affect the population density and hantavirus infection rate of rodents.9,14,34 Some studies reported that environmental variables, including temperature, may affect population dynamics of the deer mouse.14 Lower temperature in the winter may reduce the growth of rodents and the reproduction of hantavirus, and higher temperature may enable reservoirs to survive more easily in winters. Some studies showed that hantavirus infections in rodents were biased toward older animals.34,35 For example, higher hantavirus prevalence was detected in older, heavier, male deer mice in the United States.34 In addition, other studies have shown that rodent populations had a higher proportion of adults in spring–summer months than during the months of autumn and winter.35,36 As for the present study, monitoring the fluctuations of monthly maximum temperature in the area could be used as an indicator to temporally target the surveillance for HFRS infections.

Although one study reported that relative humidity may contribute to the high incidence of HFRS in northeastern China,25 another study reported that associations between HFRS epidemics and precipitation, as well as relative humidity, were not significant.37 In our study we found that relative humidity three months before analysis had a positive effect on the incidence of monthly HFRS cases, which was consistent with our previous study in northeastern China. Interestingly, we identified a negative effect of a relative humidity one-month lag on the monthly HFRS incidence. It is noteworthy that this negative effect was not found in a previous study in northeastern China.28

The model suggested that SOI with a two-month lag was negatively associated with the transmission of HFRS in Heilongjiang Province. The SOI, computed from fluctuations in the surface air pressure difference between Tahiti and Darwin, Australia, indicates development and intensity of El Niño or La Niña events in the Pacific Ocean. Sustained negative values of SOI indicate El Niño episodes, whereas sustained positive values of SOI are typical of a La Niña episode.26 It had been shown that ENSO was associated with other vector-borne diseases and zoonoses, such as malaria,38,39 dengue fever,40,41 Ross River virus infection,42,43 cutaneous leishmaniasis, and plague.44,45 Similarly, our results indicate that in addition to relative humidity and maximum temperature, SOI is also associated with HFRS incidence in Heilongjiang Province. Thus, local climate (e.g., temperature and humidity) was more likely to directly affect life cycle dynamics (e.g., reproductive rates and incubation periods) of the disease agents themselves, whereas larger-scale factors (e.g., SOI) could also influence broader ecologic processes and have possibly nonlinear impacts on disease transmission dynamics.18,46,47

A clear association between rainfall and HFRS was not detected in this study. The association between HFRS epidemics and rainfall is still a matter of debate. Some studies showed that heavy rainfall followed by increased grass seed production was associated with higher deer mouse (Peromyscus maniculatus) densities, which caused an outbreak of hantavirus pulmonary syndrome in the Four Corners region of the United States,13,16,17 and a negative relationship between rainfall and HFRS has also been reported.8 There is no clear explanation for such difference, which may reflect heterogeneity in local climate conditions. Further studies should be conducted in different regions to gain a better understanding of the effect of rainfall on HFRS transmission.

We developed a SARIMA model of HFRS incidence, which could not only quantify the relationship between the climatic variables and HFRS, but also forecast the number of HFRS cases in Heilongjiang Province. The results showed that the SARIMA model was suitable in representing variation in HFRS incidence in Heilongjiang Province during the study period because there was a good correlation between observed and predicted number of HFRS cases.

The results should be interpreted in light of the limitations of this study. First, although climate variables included in our model explain most variation in HFRS incidence in the region (as demonstrated by an adjusted R2 of 83%), more detailed independent variables could have been used. For example, human activities, farming patterns, and rodent host density also affected the transmission of HFRS.2,16 However, because of the method of data collection and retrospective nature of the study, the aforementioned factors were not available. Second, the time-series data in our study were obtained from a passive surveillance system, and some HFRS patients with milder symptoms may have been missed. Third, the study focused on Heilongjiang Province in China. This particular model in our study may not apply to the other regions in China because local climate and habitat types are different; a separate analysis may be needed for each region of concern.

In our study, we formulated and tested a climate-based HFRS epidemic forecasting SARIMA model, which uses correlations and time lags between climate factors and HFRS incidence in Heilongjiang Province. Its predictive ability has important public health implications in that it will enable local public health officials to forecast HFRS reliably and enhance HFRS prevention strategies. Although this particular model may not apply to the other regions of concern because of local variability in climate and habitat, findings of the present study will provide directions for quantitatively predicting models for supplementing early warning system in other areas, where HFRS epidemic is vulnerable to the impact of climate change. Consequently, an accurate warning system would improve public health measures in developing countries.

ACKNOWLEDGMENTS

We thank Dr. Bi Peng (University of Adelaide) and Miss Li Tian (Lund University) for valuable comments and suggestions, the Chinese Bureau of Meteorology for providing climate data, and the Heilongjiang Provincial Center for Disease Control and Prevention for providing hemorrhagic fever with renal syndrome data. Chang-Ping Li, Ricardo J. Soares Magalhaes, Ma Jun, and Wen-Yi Zhang were involved in the conceptualization, research design, execution and writing of the manuscript. Chang-Ping Li, Jun Ma, Feng-Jiang Wei, Zhuang Cui, Hai-Long Sun, and Cheng-Yi Li contributed to database design and data analysis, Bao-Long Wang, Cui Zhang, Shen-Long Li, and Liu-Yu Huang provided advice on the design of the study and the analysis and interpretation of the results. All authors were involved in the preparation of the manuscript.

  • 1.

    Ministry of Health, 1998. Handbook of Epidemic Hemorrhagic Fever Prevention and Control. Beijing: China People's Health Publishing House, 6380.

    • Search Google Scholar
    • Export Citation
  • 2.

    Schmaljohn C, Hjelle B, 1997. Hantaviruses: a global disease problem. Emerg Infect Dis 3: 95104.

  • 3.

    Bi Z, Formenty PB, Roth CE, 2008. Hantavirus infection: a review and global update. J Infect Dev Ctries 2: 323.

  • 4.

    Yan L, Fang LQ, Huang HG, Zhang LQ, Feng D, Zhao WJ, Zhang WY, Li XW, Cao WC, 2007. Landscape elements and Hantaan virus-related hemorrhagic fever with renal syndrome, People's Republic of China. Emerg Infect Dis 13: 13011306.

    • Search Google Scholar
    • Export Citation
  • 5.

    Jiang JF, Wu XM, Zuo SQ, Wang RM, Chen LQ, Wang BC, Dun Z, Zhang PH, Guo TY, Cao WC, 2006. Study on the association between Hantavirus infection and Rattus norvegicus. Chin J Epidemiol 27: 196199.

    • Search Google Scholar
    • Export Citation
  • 6.

    Zhang WY, Fang LQ, Jiang JF, Hui FM, Glass GE, Yan L, Xu YF, Zhao WJ, Yang H, Liu W, Cao WC, 2009. Predicting the risk of Hantavirus infection in Beijing, People's Republic of China. Am J Trop Med Hyg 80: 678683.

    • Search Google Scholar
    • Export Citation
  • 7.

    Bi P, Wu X, Zhang F, Parton K, Tong S, 1998. Seasonal rainfall variability, the incidence of hemorrhagic fever with renal syndrome, and prediction of the disease in low-lying areas of China. Am J Epidemiol 148: 276281.

    • Search Google Scholar
    • Export Citation
  • 8.

    Bi P, Tong S, Donald K, Parton K, Ni J, 2002. Climatic, reservoir and occupational variables and the transmission of haemorrhagic fever with renal syndrome in China. Int J Epidemiol 31: 189193.

    • Search Google Scholar
    • Export Citation
  • 9.

    Ernest SKM, Brown JH, Parmenter RR, 2000. Rodents, plants, and precipitation: spatial and temporal dynamics of consumers and resources. Oikos 88: 470482.

    • Search Google Scholar
    • Export Citation
  • 10.

    Glass GE, Shields T, Cai B, Yates TL, Parmenter R, 2007. Persistently highest risk areas for hantavirus pulmonary syndrome: potential sites for refugia. Ecol Appl 17: 129139.

    • Search Google Scholar
    • Export Citation
  • 11.

    Langlois JP, Fahrig L, Merriam G, Artsob H, 2001. Landscape structure infuences continental distribution of Hantavirus in deer mice. Landscape Ecol 16: 255266.

    • Search Google Scholar
    • Export Citation
  • 12.

    Madsen T, Shine R, 1999. Rainfall and rats: climatically-driven dynamics of a tropical rodent population. Aust J Ecol 24: 8089.

  • 13.

    Engelthaler DM, Mosley DG, Cheek JE, Levy CE, Komatsu KK, Ettestad P, Davis T, Tanda DT, Miller L, Frampton JW, Porter R, Bryan RT, 1999. Climatic and environmental patterns associated with hantavirus pulmonary syndrome, Four Corners region, United States. Emerg Infect Dis 5: 8794.

    • Search Google Scholar
    • Export Citation
  • 14.

    Luis AD, Douglass RJ, Mills JN, Bjørnstad ON, 2010. The effect of seasonality, density and climate on the population dynamics of Montana deer mice, important reservoir hosts for Sin Nombre Hantavirus. J Anim Ecol 79: 462470.

    • Search Google Scholar
    • Export Citation
  • 15.

    Glass GE, Cheek JE, Patz JA, Shields TM, Doyle TJ, Thoroughman DA, Hunt DK, Enscore RE, Gage KL, Irland C, Peters CJ, Bryan R, 2000. Using remotely sensed data to identify areas at risk for Hantavirus pulmonary syndrome. Emerg Infect Dis 6: 238247.

    • Search Google Scholar
    • Export Citation
  • 16.

    Hjelle B, Glass GE, 2000. Outbreak of Hantavirus infection in the Four Corners region of the United States in the wake of the 1997–1998 El Niño-Southern Oscillation. J Infect Dis 181: 15691573.

    • Search Google Scholar
    • Export Citation
  • 17.

    Tamerius JD, Wise EK, Uejio CK, McCoy AL, Comrie AC, 2007. Climate and human health: synthesizing environmental complexity and uncertainty. Stochastic Environ Res Risk Assess 21: 601613.

    • Search Google Scholar
    • Export Citation
  • 18.

    Tersago K, Verhagen R, Servais A, Heyman P, Ducoffre G, Leirs H, 2009. Hantavirus disease (nephropathia epidemica) in Belgium: effects of tree seed production and climate. Epidemiol Infect 137: 250256.

    • Search Google Scholar
    • Export Citation
  • 19.

    Klempa B, 2009. Hantaviruses and climate change. Clin Microbiol Infect 15: 518523.

  • 20.

    Bi P, Parton K, 2003. El Niño and incidence of hemorrhagic fever with renal syndrome in China. JAMA 289: 176177.

  • 21.

    Fang LQ, Wang XJ, Liang S, Yan LL, Song SX, Zhang WY, Qian Q, Li YP, Wei L, Wang ZQ, Yang H, Cao WC, 2010. Spatiotemporal trends and climatic factors of hemorrhagic fever with renal syndrome epidemic in Shandong Province, China. PLoS Negl Trop Dis 4: 110.

    • Search Google Scholar
    • Export Citation
  • 22.

    Pettersson L, Boman J, Juto P, Evander M, Ahlm C, 2008. Outbreak of Puumala virus infection, Sweden. Emerg Infect Dis 14: 808810.

  • 23.

    Schwarz AC, Ranft U, Piechotowski I, Childs JE, Brockmann SO, 2009. Risk factors for human infection with Puumala virus, southwestern Germany. Emerg Infect Dis 15: 10321039.

    • Search Google Scholar
    • Export Citation
  • 24.

    Huang RH, Wu YF, 1989. The infuence of ENSO on the summer climate change in China and its mechanism. Adv Atmos Sci 6: 2132.

  • 25.

    Zhang WY, Guo WD, Fang LQ, Li CP, Bi P, Glass GE, Jiang JF, Sun SH, Qian Q, Liu W, Yan L, Yang H, Tong SL, Cao WC, 2010. Climate variability and hemorrhagic fever with renal syndrome transmission in northeastern China. Environ Health Perspect 118: 915920.

    • Search Google Scholar
    • Export Citation
  • 26.

    National Centers for Environmental Prediction, 2010. Climatic Forecast System. Available at: http://www.cpc.ncep.noaa.gov/data/indices/soi. Accessed June 6, 2010.

    • Search Google Scholar
    • Export Citation
  • 27.

    Hu W, Tong S, Mengersen K, Connell D, 2007. Weather variability and the incidence of cryptosporidiosis: comparison of time series poisson regression and SARIMA models. Ann Epidemiol 17: 679688.

    • Search Google Scholar
    • Export Citation
  • 28.

    Allard R, 1998. Use of time-series analysis in infectious disease surveillance. Bull World Health Organ 76: 327333.

  • 29.

    Box G, Jenkins G, 1970. Time-Series Analysis: Forecasting and Control. San Francisco, CA: Holden-Day, 376382.

  • 30.

    Hu W, Mengersen K, Bi P, Tong S, 2007. Time series analysis of the risk factors for hemorrhagic fever with renal syndrome: comparison of statistical models. Epidemiol Infect 135: 245252.

    • Search Google Scholar
    • Export Citation
  • 31.

    Bozdogan H, 1987. Model-selection and Akaike's information criterion (AIC): the general theory and its analytical extensions. Psychometrika 52: 345370.

    • Search Google Scholar
    • Export Citation
  • 32.

    Liddle AR, 2007. Information criteria for astrophysical model selection. Mon Not R Astron Soc 377: 7478.

  • 33.

    SAS, 2008. SAS/ETS®9.2: User's Guide. Cary, NC: SAS Institute Inc., 177300.

  • 34.

    Calisher CH, Wagoner KD, Amman BR, Root JJ, Douglass RJ, Kuenzi AJ, Abbott KD, Parmenter C, Yates TL, Ksiazek TG, Beaty BJ, Mills JN, 2007. Demographic factors associated with prevalence of antibody to Sin Nombre virus in deer mice in the western United States. J Wildl Dis 43: 111.

    • Search Google Scholar
    • Export Citation
  • 35.

    Glass GE, Livingstone W, Mills JN, Hlady WG, Fine JB, Biggler W, Coke T, Frazier D, Atherley S, Rollin PE, Ksiazek TG, Peters CJ, Childs JE, 1998. Black Creek Canal virus infection in Sigmodon hispidus in southern Florida. Am J Trop Med Hyg 59: 699703.

    • Search Google Scholar
    • Export Citation
  • 36.

    Yahnke CJ, Meserve PL, Ksiazek TG, Mills JN. Patterns of infection with Laguna Negra virus in wild populations of Calomys laucha in the central Paraguayan Chaco 2001. Am J Trop Med Hyg 65: 768776.

    • Search Google Scholar
    • Export Citation
  • 37.

    Luo CW, Liu QY, Hou JL, 2009. Correlation analysis and regression model of epidemic factors of hemorrhagic fever with renal syndrome in Heihe City, Heilongjiang Province. Dis Surveill 24: 118120.

    • Search Google Scholar
    • Export Citation
  • 38.

    Thomson MC, Mason SJ, Phindela T, Connor SJ, 2005. Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana. Am J Trop Med Hyg 73: 214221.

    • Search Google Scholar
    • Export Citation
  • 39.

    Zhou G, Minakawa N, Githeko AK, Yan G, 2004. Association between climate variability and malaria epidemics in the east African highlands. Proc Natl Acad Sci USA 101: 23752380.

    • Search Google Scholar
    • Export Citation
  • 40.

    Cazelles B, Chavez M, McMichael AJ, Hales S, 2005. Nonstationary influence of El Niño on the synchronous dengue epidemics in Thailand. PLoS Med 2: e106.

    • Search Google Scholar
    • Export Citation
  • 41.

    Hales S, de Wet N, Maindonald J, Woodward A, 2002. Potential effect of population and climate changes on global distribution of dengue fever: an empirical model. Lancet 360: 830834.

    • Search Google Scholar
    • Export Citation
  • 42.

    Tong S, Hu W, 2001. Climate variation and incidence of Ross river virus in Cairns, Australia: a time-series analysis. Environ Health Perspect 109: 12711273.

    • Search Google Scholar
    • Export Citation
  • 43.

    Tong S, Hu W, 2002. Different responses of Ross River virus to climate variability between coastline and inland cities in Queensland, Australia. Occup Environ Med 59: 739744.

    • Search Google Scholar
    • Export Citation
  • 44.

    Chaves LF, Pascual M, 2006. Climate cycles and forecasts of cutaneous leishmaniasis, a nonstationary vector-borne disease. PLoS Med 3: 295.

  • 45.

    Nakazawa Y, Williams R, Peterson AT, Mead P, Staples E, Gage KL, 2007. Climate change effects on plague and tularemia in the United States. Vector Borne Zoonotic Dis 7: 529540.

    • Search Google Scholar
    • Export Citation
  • 46.

    Hallett TB, Coulson T, Pilkington JG, Clutton-Brock TH, Pemberton JM, Grenfell BT, 2004. Why large-scale climate indices seem to predict ecological processes better than local weather. Nature 430: 7175.

    • Search Google Scholar
    • Export Citation
  • 47.

    Zheng ZM, Jiang ZK, Chen AG, 2008. Rodents Zoology. Shanghai: Shanghai Jiaotong University Press, 400450.

Author Notes

* Address correspondence to Wen-Yi Zhang, Institute of Disease Control and Prevention of PLA, 20 Dong-Da Street, Fengtai District, Beijing, 100071, People's Republic of China, E-mail: zwy0419@126.com, or Jun Ma, Department of Health Statistics, College of Public Health, Tianjin Medical University, Tianjin, 300070, People's Republic of China, E-mail: majun@tijmu.edu.cn.† These authors contributed equally to this article.

Financial support: This study was supported by the grants from National Natural Science Foundation of China (81102169, 21003077), the National Basic Research Program of China (973 Program) (2012CB955500-955504), and the Open Project of Key Laboratory of Advanced Energy Materials Chemistry College Nankai University (KLAEMC-OP201201).

Authors' addresses: Chang-Ping Li, Zhuang Cui, Bao-Long Wang, and Jun Ma, College of Public Health, Tianjin Medical University, Heping District, Tianjin, China, E-mails: changpingli@163.com, cuizhuang@tijmu.edu.cn, lg_20o3@sina.com, and majun@tijmu.edu.cn. Shen-Long Li, Hai-Long Sun, Cheng-Yi Li, Liu-Yu Huang, and Wen-Yi Zhang, Institute of Disease Control and Prevention, Academy of Military Medical Science, Fengtai District, Beijing, China, E-mails: lishenlong@sohu.com, xxsunhl@yahoo.com.cn, licy_60@163.com, huangly@nic.bmi.ac.cn, and zwy0419@126.com. Ricardo J. Soares Magalhaes, School of Population Health, University of Queensland, Herston, Queensland, Australia. Cui Zhang, Nankai University, Nankai District, Tianjin, China, E-mail: r.magalhaes@uq.edu.au.

Save