Background
Hemorrhagic fever with renal syndrome (HFRS) is a zoonosis caused by Hantaan or Hantaan-related virus, with characteristics of fever, hemorrhage, kidney damage and hypotension. In China, HFRS was recognized in northeastern China in 1931[
1]. There is a concern recently for an outbreak in some countries [
2‐
4]. China is the most severe endemic country, with 90% of the total HFRS worldwide cases reported[
1]. Although vaccination, rodent control, environment management and other related measures have been implemented, HFRS remains a serious public health problem in mainland China with 20,000–50,000 human cases annually reported[
5]. Liaoning Province is one of the most serious affected areas with the most cases in mainland China and the highest incidence during the years 2004 and 2005[
6], and an outbreak of HFRS was also reported in Liaoning in 2006[
7]. There are two kinds of rodent-borne virus epidemic of HFRS in Liaoning province: Hantaan virus transmitted by the striped field mouse (
Apodemus agrarius) and Seoul virus transmitted by the brown rat (
Rattus norvegicus). Although the incidence of HFRS is stable and has a descending trend at the national level in China, the scope of epidemic focus of the Seoul virus is still expanding. Huludao city was selected as the study area because Huludao city is the traditional HFRS epidemic focus of the Seoul virus. It was recognized as an epidemic focus of HFRS in 1984 and has since served as a national HFRS surveillance site[
8,
9].
It has been accepted that climate plays a role in the transmission of many infectious diseases including HFRS [
10‐
14]. Several studies have explored the association between HFRS and climate variation [
15‐
19]. The climatic variables, such as temperature, not only affect the rate of replication of virus, but also have an impact on the environmental reservoirs-rodents [
20‐
22]. Thus, the climatic variables are expected to influence the incidence of HFRS through rodents. Because the variation of HFRS incidence are linked inherently to the climate and reservoir factors, to what degree the relationship between climatic variables, reservoir information and HFRS is a focus of interest in the present study.
Correlation and regression analysis are common methods for existing studies on the transmission of HFRS. The limitations of these methods are that some variables, like climate and reservoir, can't be described clearly and the interaction between them could not be explored. Structural equation model (SEM)[
23,
24] is a statistical technique for testing and estimating causal relationships using a combination of statistical data and qualitative causal assumptions. SEMs were originally developed in the early 1970s in the field of social science of fit models with variables that cannot be measured or observed. A key feature of the SEM approach is that it allows one to compare candidate models. SEM grows out of and serves purposes similar to multiple regression. It is a combination of multiple regression and factor analysis. Recently, SEM has become a particularly attractive data-analytic option because of the development of several new types of models and software capabilities that are particularly well suited to the research interests of clinical scientists and public health managers[
25,
26].
In the present study, SEM was adopted to analyze the influence of climate, reservoir on the incidence of HFRS and describe the effect in quantity. It was expected that HFRS forecasts or control decision-making could be based on the relationships between HFRS and correlated climatic and reservoir factors in both surveyed and similar unsurveyed areas.
Discussion
The HFRS incidence increased from 1990 to 2006 in the study area. The reasons for the increase of HFRS incidence may be as follows: Firstly, more and more urban constructions were built and rebuilt in recent years. Rapid development of China's urban construction resulted in the frequent migration of R. norvegicus and more opportunities on contact with human being. Secondly, the increase of rural migrant workers resulted in the increase of susceptible population. Thirdly, the fundamental reason was that the density of rodents went upward with a high virus-carrying rate recently.
Epidemics of HFRS can fall with a well-defined periodicity and/or seasonality[
38]. This variation is important because it can have significant implications for the design and effectiveness of control strategies. There is a clear seasonal variation of HFRS incidence in Huludao City and the incidence increased with a 2 to 3 years cycle during the study period. In China, at the national level, there was a peak of HFRS incidence every 8 years, at the county or city level, the cycle were 3 to 5 years[
1]. The possible reason for the 2–3 year cycle in Huludao may be that the density of rodents and virus-carrying rate among rodents were at a high level during the study period. The density of rodents ranged from 1% to 12% with an average 5%, and virus-carrying rate among rodents ranged from 4% to 14% with an average 9%[
33].
In 2005, the World Health Organization report on climate and prediction of infectious disease epidemics[
39] indicated that the construction of weather- and climate-based systems to provide early warning of incipient epidemics was now feasible and could provide considerable population health benefit if informational, structural, and monetary barriers to implementation could be overcome. Several studies indicated the effect of climatic or reservoir variables on the incidence of HFRS, but few studies described the effect in quantity. Liaoning Province can be divided into three parts in respect to terrain: the Liaodong mountainous region, the Liaohe plain region, and the Liaoxi mountainous region[
40]. Huludao City lies in the Liaoxi mountainous region. Lin H et al indicated that the terrain, the amount of forestation and the relative high humidity might be the important factors responsible for the epidemic development of HFRS in Liaoning Province, but they didn't explain what the association really was[
41]. Our findings about the relationship between HFRS incidence, climatic factors and reservoir are of part concordance with the results of other studies in China[
16,
42,
43]. Liu et al applied case-crossover design and conditional logistic regression to analyze the relationship between meteorological factors and HFRS incidence in one national surveillance spot of HFRS in Shandong province[
42]. They found that mean temperature, rainfall, humidity sunlight, air pressure and velocity of wind were associated with HFRS incidence. Based on the surveillance data in Jiangsu Province, Wu et al indicated that meteorological factors could be used to predict HFRS incidence[
43]. However, the climatic variables' effects on HFRS incidence via rodents were not considered in these studies.
In the present study, we discussed how climatic factors acted on reservoir and influenced the incidence of HFRS in the end. Because the climatic variables are not independent variables, the regression coefficients are very unstable with excessive standard error. We presented an SEM implementation to solve this problem. The final SEM indicated that climate affected HFRS incidence mainly through the effect on reservoir in the study area, reservoir had a greater effect in the model. The goodness of fit statistics meant that SEM was suitable for understanding such relational data in multivariate systems. The latent variable "climate" here refers to the average state over a longer time period[
44]. SEM that had climatic variable with lagged HFRS incidence was designed because it allowed the examination of a lagged effect of climate variability to impact on the incidence of HFRS. The lag would capture the period of rodents growth, virus development time within the rodents and the virus incubation period within the human body[
16]. The SEMs approach also has several limitations, such as 1) the idea that sample covariance instead of the sample values themselves was used to fit by modeling may destroy valuable data; 2) as the number of explanatory increases, the number of parameters in the model can increase exponentially; 3) SEM is usually viewed as a confirmatory rather than exploratory procedure, can't advance new casual model by itself.
As a national HFRS surveillance site, good records about HFRS incidence have been kept in Huludao City. The blood samples of HFRS cases were collected in the hospitals, serologic identification was then performed at the laboratory to confirm the clinical diagnosis. There might be admission rate bias in the disease report, but this has been reduced as much as possible[
41]. The Chinese Government established a routine reporting system for selected infectious diseases in the 1950s, and the system switched from paper-based reporting to the submission of electronic files in 1985, and since 2003 has used web-based reporting[
45]. Under-reporting is possible in the disease surveillance system, especially before the computer and web were widely applied. An investigation of missing reports of notifiable diseases in China in 2005 indicated that there was much to be improved upon in Chinese medical facilities as far as reporting of infectious diseases[
46]. So, the relationship between climatic variables, reservoir information and HFRS may be underestimated. Further, in the present study only the new symptomatic HFRS cases were used to calculate the HFRS incidence, while, the majority of human hantavirus infections are asymptomatic[
47]. In China, Song G reported that relatively high inapparent infection rates(8%–20%) in the population of endemic areas of the Rattus-type HFRS after big outbreaks played a significant role in the gradual decline of the incidence of HFRS [
48], Bi P reported that the antibody tire of asymptomatic infection was too low to prevent the development of clinical cases[
49]. The asymptomatic infection rate among healthy population in the study area varied from 3.6% to 9.2% with an average 5%, which may serve as a potential limitation of the study. Another limitation of the present study may be that we only focused on the relationship between the meteorological factors, reservoir factors and the incidence of HFRS. The occupational activities, such as farming and mining, may affect people's contact with rodents[
16]. However, in epidemic foci of the Seoul virus, people who take different jobs have similar chance of contacting with
Rattus norvegicus [
1].
Acknowledgements
This study was financially supported by the National Natural Science Foundation of China(NNSFC), No. 70503028 and No. 30771860. Peng Guan was supported by the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry, No. [2008]890 and a CMU Development grant, No. [2008]5. The authors are grateful to all the participants in this study. The authors thank Bo Qu and Haiqiang Guo for their helpful discussion with data analyses. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official views of the NNSFC. The authors thank the editors and three reviewers for their insightful comments which led to a significantly improved version of the manuscript. Also, the authors thank Dr. Weihua Zhang (Professor of Liberal Arts, Savannah College of Art and Design, Savannah, Georgia, USA) for her careful checking of grammar and spelling of the manuscript.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PG conceived the study and drafted the manuscript. DH, PG and BZ secured funding. PG, DH and MH managed and analyzed the data. BZ contributed to the analysis design, and reviewed the manuscript. TS contributed to the HFRS incidence and reservoir data collection and managed the HFRS incidence database which was administered and supervised by JG. TS and MH contributed to the results interpretation. All authors contributed to the writing of the final version of this paper.