This article has Open Peer Review reports available.
Ecology and geography of hemorrhagic fever with renal syndrome in Changsha, China
- Hong Xiao†1Email author,
- Xiaoling Lin†1,
- Lidong Gao†2,
- Cunrui Huang3,
- Huaiyu Tian1Email author,
- Na Li4,
- Jianxin Qin1,
- Peijuan Zhu1,
- Biyun Chen2,
- Xixing Zhang5 and
- Jian Zhao6
© Xiao et al.; licensee BioMed Central Ltd. 2013
Received: 24 February 2012
Accepted: 17 June 2013
Published: 3 July 2013
Hemorrhagic fever with renal syndrome (HFRS) is an important public health problem in mainland China. HFRS is particularly endemic in Changsha, the capital city of Hunan Province, with one of the highest incidences in China. The occurrence of HFRS is influenced by environmental factors. However, few studies have examined the relationship between environmental variation (such as land use changes and climate variations), rodents and HFRS occurrence. The purpose of this study is to predict the distribution of HFRS and identify the risk factors and relationship between HFRS occurrence and rodent hosts, combining ecological modeling with the Markov chain Monte Carlo approach.
Ecological niche models (ENMs) were used to evaluate potential geographic distributions of rodent species by reconstructing details of their ecological niches in ecological dimensions, and projecting the results onto geography. The Genetic Algorithm for Rule-set Production was used to produce ENMs. Data were collected on HFRS cases in Changsha from 2005 to 2009, as well as national land survey data, surveillance data of rodents, meteorological data and normalized difference vegetation index (NDVI).
The highest occurrence of HFRS was in districts with strong temperature seasonality, where elevation is below 200 m, mean annual temperature is around 17.5°C, and annual precipitation is below 1600 mm. Cultivated and urban lands in particular are associated with HFRS occurrence. Monthly NDVI values of areas predicted present is lower than areas predicted absent, with high seasonal variation. The number of HFRS cases was correlated with rodent density, and the incidence of HFRS cases in urban and forest areas was mainly associated with the density of Rattus norvegicus and Apodemus agrarius, respectively.
Heterogeneity between different areas shows that HFRS occurrence is affected by the intensity of human activity, climate conditions, and landscape elements. Rodent density and species composition have significant impacts on the number of HFRS cases and their distribution.
Hemorrhagic fever with renal syndrome (HFRS) is a group of rodent-borne diseases caused by hantaviruses (HV). It is an important public health problem in China, with 30,000 to 60,000 cases reported annually, accounting for 90% of human cases reported globally over the last 10 years . In China, HFRS is caused mainly by two types of hantavirus, Hantaan virus (HTNV) and Seoul virus (SEOV), each of which has co-evolved with a distinct rodent host . HTNV is carried by Apodemus agrarius, and SEOV by Rattus norvegicus. Humans usually become infected with hantaviruses through contact with or inhalation of aerosols and secretions from infected rodent hosts.
Hunan Province has become one of the most severely endemic areas in China since the first case was discovered in 1963. It is a traditional and mixed epidemic area, with SEOV the main hantavirus type and Mus musculus, R. norvegicus and A. agrarius the dominant rodent species. The incidence rate of HFRS peaked at 101.68/100,000 in 1994, and it has become a great threat to the health of the local population. Previous studies have found that precipitation, temperature, humidity, normalized difference vegetation index (NDVI), and land use types are important risk factors for HFRS incidence [3–6].
In recent years, ecological niche models (ENMs) have been developed and used to examine associations between spatial risk factors and ecological niches in disease transmission . Species distributions are constrained by a series of evolutionary adaptations that are generally conceptualized as the ecological niche. Disease transmission involves interacting species (pathogens, vectors and hosts), and the conjunction of their individual ecological niches determines patterns of transmission. ENMs can be used to analyze geographic and ecological information, and then quantify potential risks based on vector, host and pathogen data [8, 9]. To characterize geographic patterns of disease transmission, traditional methods usually summarize overall patterns and trends in the form of a smoothed surface, at some level of generalization or averaging. This may involve loss of resolution and may not take into account the fine-scale ecological variation that underlies transmission patterns. However, ENMs permit fine-scale characterization of geographic patterns without loss of resolution [7, 10].
The characteristics of HFRS vary with locality, owing to the vast and complex topography, intricate ecological environments, and varied climates of China. It is important to identify the risk factors of HFRS and predict high risk areas, thereby facilitating application of prevention strategies to avoid or decrease loss of life. Since rodent population density and species composition have significant impacts on HFRS occurrence, it is of great importance to analyze the relationship between rodent host and HFRS incidence. In this study, we investigated the effects of ecological and geographic factors on HFRS occurrence using ENMs, and explored the relationship between HFRS cases and rodent hosts by the Markov chain Monte Carlo (MCMC) method, using data from Changsha, China.
The study protocol was reviewed by the institutional review board of the Hunan CDC and ethics approval was not required.
Data collection and management
Surveillance on hantavirus infections among rodent hosts was carried out in Ningxiang County, according to the protocol established by Chinese Centers for Disease Control and Prevention. Representative villages were selected for rodent surveillance, according to the distribution of HFRS epidemics and landscape elements. Indoor and outdoor rodent traps with peanuts were set at night, and recovered in the morning. The traps were placed about 1 km away from villages in locations where rodents were most likely found, such as the edges of rivers and roads, on ridges and in yards. Traps were set in spring (March to April) and autumn (September to October). More than 100 traps per patch were placed indoors at approximately 12–15-meter intervals for three consecutive nights, and more than 200 traps per patch were placed outdoors (every 5 meters in each row, with 50 meters between rows). A total of 22,060 traps were set, and 728 rodents were captured from 2005 to 2009. Relative population density of rodents was used as an indicator of abundance. This is calculated as the number of rodents captured, divided by the number of traps.
Ecogeographical variables for modeling
Derived from elevation
Derived from elevation
The Second National Land Survey Data
United States Geological Survey (http://eros.usgs.gov)
Human Footprint index
International Earth Science Information Center(http://www.ciesin.columbia.edu/)
Changsha Statistical Yearbook
World Wildlife Fund(http://conserveonline.org)
National Aeronautics and Space Administration (http://ladsweb.nascom.nasa.gov/data/)
Ecological niche models
Test methods of ecological niche models
(a + d) / (a + b + c + d)
c / (a + c)
b / (b + d)
To build different subset models for the entire occurrence area, an algorithm threshold of 0.01 was selected, with 1,000 iterations as an upper limit for each replicate. Because of the stochastic nature of GARP in producing different outputs at different replicates, best practice approaches were required . Ten best subsets were selected, with a threshold level of 0% extrinsic hard omission and 50% commission. GARP outputs are rasterized coverages of the study area. Two different pixel values were used to show the absence or presence of related species; 0 means absence (including areas in which no rules can be applied) and 1 means presence. Then, we calculated each pixel value by summing the "best subset models", using the Raster Calculator function in the Spatial Analyst extension of ArcGIS 9.3, producing final predictions of potential distributions with 11 thresholds (integers from 0 to 10). A pixel was predicted to be present if the threshold ≥ 5 of the 10 best subset models.
We developed a series of tests of model predictive ability. In each case, the developed models and predictions tested were based on independent suites of HFRS data. The HFRS data were divided into quadrants, above and below the median longitude and median latitude: (i) west versus east of the median longitude (hereafter “WE” tests); (ii) north versus south of the median latitude (hereafter “NS” tests); (iii) on-diagonal (upper left-hand and lower right-hand quadrants) and off-diagonal (upper right-hand and lower left-hand quadrants, hereafter “DIAG” tests). In each case, we developed both reciprocal predictions, testing the ability of ENMs to predict the distribution of HFRS in regions where no sampling was available. Finally, data from 2005 through 2008 were used to develop the ENMs, and 2009 data were used to validate the prediction (hereafter “Time” tests).
To investigate relative contributions of the 25 environmental variables, a jackknife procedure was used, which was performed by fitting the model with n-1 variables and successively omitting one variable. Then, we compared the predictive accuracy of the subset models with the comprehensive model (based on n variables), to eliminate variables that lead to over-fitting and to select key risk factors. Occurrence data from 2005 through 2008 were used to perform the jackknife procedure, and the average area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate predictive accuracy of the model. After the jackknife procedure, the final HFRS prediction was executed again by GARP, using available variables.
Rodent species composition and HFRS occurrence
Modeling and testing approaches
Summary of model predictions and tests for Changsha
No. of successes
Proportion of predicted area
Cumulative bino-mial probability
North predicts south
South predicts north
West predicts east
East predicts west
On predicts off
Off predicts on
05-08 predicts 09
Summary of statistical analysis of jackknife procedure
Human footprint index
Ecological characteristics of predicted results
Rodent host and HFRS occurrence
The coefficient matrix implies that HFRS cases in urban areas and cultivated land are mainly caused by R. norvegicus and other species (including Rattus niviventer, R. flavipectus, Microtus fortis and shrew species, with total proportion less than 5%), and that risk in forested areas is from A. agrarius. In terms of population quantity, R. norvegicus and A. agrarius are considered the dominant species, with varying impacts on HFRS occurrence in different risk areas of Changsha.
With rapid development of the economy and urbanization, the environment in China has changed greatly. Continuous habitat changes for rodent hosts are believed to influence the spatial distribution of HFRS, extending the main epidemic areas from rural to urban areas. This study improves our understanding of the transmission patterns of HFRS. The correlation between rodent species composition and HFRS occurrence also provides a valuable approach for predicting HFRS risk.
Human activities including rapid urbanization, deforestation, agricultural invasion, land use change, pollution and population migration are considered important factors for the outbreak and reemergence of various infectious diseases, and for heterogeneity in the incidence and spatial distribution of these diseases [17–19]. Forest is the most extensive land cover type in Changsha, but the risk level there is low. However, risk levels in cultivated land and urban areas are relatively high.
NDVI is correlated with amount and productivity of vegetation and crops, which are the main food sources for rodents. The NDVI is higher in thick vegetative areas, such as forests and orchards, and lower in cultivated land areas, with significant seasonal variations. HFRS occurrence is closely associated with NDVI. HFRS risk is also affected by differences in local climates. Meteorological factors, including humidity, temperature and precipitation, influence not only the infectivity and vitality of hantavirus but also the distribution of rodent hosts . Moist environments are conducive to the vitality and infectivity of Hantaan virus and to the existence and distribution of rodents. However, too much precipitation destroys the rodent habitat, decreasing the population . HFRS cases are rarely reported in very dry or very wet areas . In Changsha, the high risk areas mostly concentrate in locations with annual temperature around 17.5°C and annual precipitation between 1,300 and 1,600 mm.
The density and species composition of rodent hosts are closely associated with HFRS occurrence. Usually, HFRS incidence escalates with rodent density, and vice versa. Each hantavirus species is predominately associated with a distinct or a few related rodents as its host. Since different rodent species adapt to various environments and habitats, rodent impacts vary by area. This study showed that A. agrarius and R. norvegicus are the two predominant animal hosts; A. agrarius is abundant in forested areas, and R. norvegicus is predominant in urban and cultivated areas. Correlation analysis between HFRS incidence and rodent species composition indicates that the main HFRS risks are concentrated in urban and cultivated land and forests, which is consistent with ENMs prediction.
ENMs were used to identify the relationship between HFRS cases and the environment. This was done to explore non-random association between environmental characteristics and known epidemic areas and the entire study area, and to predict potential risks based on the basic ecological demands of HFRS. The MCMC method was used to predict HFRS cases in various areas based on surveillance data of rodent hosts, and cases in diverse land use types. The MCMC results further validated the predictive results of the ENMs. Therefore, this study offers a valuable approach to predict HFRS risk and to analyze transmission patterns in cases when “absence” data are unavailable. The results have important implications for the prevention and control of HFRS. The final prediction model can be used as a guide map for possible future outbreaks in Changsha, and related risk factors may be used as predictors of HFRS occurrence. For the large central area of the map (Changsha County), a predicted high risk area, special measures may needed to prevent the spread of HFRS. Since A. agrarius is abundant in forests, and R. norvegicus is predominant in urban and cultivated areas, different preventive measures are needed in different areas. All this information provides guidance for public health action.
The results of temporal testing based on the entire region are clearly better than other spatial calibration tests, and only a few test points were correctly predicted in the WE tests (especially the "east predicts west" test). More applications may be required to investigate ecological differences between HFRS occurrence in the eastern and western areas of Changsha. For optimal performance of temporal testing, time-specific ENMs are suggested for obtaining more specific and detailed disease occurrence information. Limitations of this study should also be acknowledged. The data are from a passive surveillance system, so their quality is not as high as those collected from active surveillance. The environmental variables used in the models are nearly abiotic factors, but in fact, biotic factors are important in disease occurrence. Previous studies have indicated that varying modeling approaches can yield substantially different predictions [21, 22]; therefore, selection of modeling methods and environmental variables should be taken into careful consideration. Future work is required to address these issues.
The authors extend their appreciation to Ruchun Liu and Tianmu Chen for providing the data. This work was supported by the Key Discipline Construction Project in Hunan Province (2008001), National Natural Science Foundation (40971038), Scientific Research Fund of the Hunan Provincial Education Department (11K037), Hunan Provincial Natural Science Foundation of China (11JJ3119), Science and Technology Planning Project of Hunan Province, China (2010SK3007), and Key Subject Construction Project of Hunan Normal University (geographic information systems).
- Luo CW, Chen HX: Study on the factors influenced epidemic of hemorrhagic fever with renal syndrome [in Chinese]. Chin J Vector Bio & Control. 2003, 14: 451-454.Google Scholar
- Papa A, Bojovic B, Antoniadis A: Hantaviruses in Serbia and Montenegro. Emerg Infect Dis. 2006, 12: 1015-1018. 10.3201/eid1206.051564.View ArticlePubMedPubMed CentralGoogle Scholar
- Zhang WY, Guo WD, Fang LQ, Li CP, Bi P, Glass GE, Jiang JF, Sun SH, Qian Q, Liu W, et al: Climate variability and hemorrhagic fever with renal syndrome transmission in northeastern China. Environ Health Persp. 2010, 118: 915-920. 10.1289/ehp.0901504.View ArticleGoogle Scholar
- Fang LQ, Wang XJ, Liang S, Li YL, Song SX, Zhang WY, Qian Q, Li YP, Wei L, Wang ZQ, et al: Spatiotemporal trends and climatic factors of hemorrhagic fever with renal syndrome epidemic in Shandong Province. China. PLoS Negl Trop Dis. 2010, 4: e789-10.1371/journal.pntd.0000789.View ArticlePubMedGoogle Scholar
- Viel JF, Lefebvre A, Marianneau P, Joly D, Giraudoux P, Upegui E, Tordo N, Hoen B: Environmental risk factors for haemorrhagic fever with renal syndrome in a French new epidemic area. Epidemiol Infect. 2011, 139: 867-874. 10.1017/S0950268810002062.View ArticlePubMedGoogle Scholar
- Yan L, Fang LQ, Huang HG, Zhang LQ, Feng D, Zhao WJ, Zhang WY, Li XW, Cao WC: Landscape elements and Hantaan virus-related hemorrhagic fever with renal syndrome, People's Republic of China. Emerg Infect Dis. 2007, 13: 1301-1306. 10.3201/eid1309.061481.View ArticlePubMedPubMed CentralGoogle Scholar
- Peterson AT: Ecologic niche modeling and spatial patterns of disease transmission. Emerg Infect Dis. 2006, 12: 1822-1826. 10.3201/eid1212.060373.View ArticlePubMedPubMed CentralGoogle Scholar
- Guisan A, Zimmermann NE: Predictive habitat distribution models in ecology. Ecol Model. 2000, 135: 147-186. 10.1016/S0304-3800(00)00354-9.View ArticleGoogle Scholar
- Wei L, Qian Q, Wang ZQ, Glass GE, Song SX, Zhang WY, Li XJ, Yang H, Wang XJ, Fang LQ, et al: Using Geographic Information System-based Ecologic Niche Models to Forecast the Risk of Hantavirus Infection in Shandong Province, China. Am J Trop Med Hyg. 2011, 84: 497-503. 10.4269/ajtmh.2011.10-0314.View ArticlePubMedPubMed CentralGoogle Scholar
- Peterson AT: Ecological niche modelling and understanding the geography of disease transmission. Vet Ital. 2007, 43: 393-400.PubMedGoogle Scholar
- Ministry of Health: Handbook of Epidemic Hemorrhagic Fever Prevention and Control. 1998, Beijing: China People's Health Publishing HouseGoogle Scholar
- Stockman AK, Beamer DA, Bond JE: An evaluation of a GARP model as an approach to predicting the spatial distribution of non‒vagile invertebrate species. Divers Distrib. 2006, 12: 81-89. 10.1111/j.1366-9516.2006.00225.x.View ArticleGoogle Scholar
- Stockwell D, Peters D: The GARP modelling system: problems and solutions to automated spatial prediction. Int J Geogr Inf Sci. 1999, 13: 143-158. 10.1080/136588199241391.View ArticleGoogle Scholar
- Anderson RP, Lew D, Peterson AT: Evaluating predictive models of species' distributions: criteria for selecting optimal models. Ecol Model. 2003, 162: 211-232. 10.1016/S0304-3800(02)00349-6.View ArticleGoogle Scholar
- Phillips SJ, Anderson RP, Schapire RE: Maximum entropy modeling of species geographic distributions. Ecol Model. 2006, 190: 231-259. 10.1016/j.ecolmodel.2005.03.026.View ArticleGoogle Scholar
- Haario H, Laine M, Mira A, Saksman E: DRAM: efficient adaptive MCMC. Stat Comput. 2006, 16: 339-354. 10.1007/s11222-006-9438-0.View ArticleGoogle Scholar
- Pimentel D, Cooperstein S, Randell H, Filiberto D, Sorrentino S, Kaye B, Nicklin C, Yagi J, Brian J, O’hern J, et al: Ecology of increasing diseases: population growth and environmental degradation. Hum Ecol. 2007, 35: 653-668. 10.1007/s10745-007-9128-3.View ArticleGoogle Scholar
- Patz JA, Daszak P, Tabor GM, Aguirre AA, Pearl M, Epstein J, Wolfe ND, Kilpatrick AM, Foufopoulos J, Molyneux D, et al: Unhealthy landscapes: policy recommendations on land use change and infectious disease emergence. Environ Health Persp. 2004, 112: 1092-1098. 10.1289/ehp.6877.View ArticleGoogle Scholar
- Fang LQ, Yan L, Liang S, De Vlas SJ, Feng D, Han X, Zhao W, Xu B, Bian L, Yang H, et al: Spatial analysis of hemorrhagic fever with renal syndrome in China. BMC Infect Dis. 2006, 6: 77-10.1186/1471-2334-6-77.View ArticlePubMedPubMed CentralGoogle Scholar
- Bi P, Wu X, Zhang F, Parton KA, Tong S: 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. 1998, 148: 276-281. 10.1093/oxfordjournals.aje.a009636.View ArticlePubMedGoogle Scholar
- Elith J, Graham CH, Anderson RP, Dudik M, Ferrier S, Guisan A, Hijmans RJ, Huettmann F, Leathwick JR, Lehmann A, et al: Novel methods improve prediction of species' distributions from occurrence data. Ecography. 2006, 29: 129-151. 10.1111/j.2006.0906-7590.04596.x.View ArticleGoogle Scholar
- Tsoar A, Allouche O, Steinitz O, Rotem D, Kadmon R: A comparative evaluation of presence-only methods for modelling species distribution. Divers Distrib. 2007, 13: 397-405. 10.1111/j.1472-4642.2007.00346.x.View ArticleGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2334/13/305/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.