|BECKLEY, C - University Of Ghana|
|SHABAN, S - University Of Ghana|
|KOKU, R - University Of Ghana|
|PALMER, G - University Of Washington|
|HUDAK, A - Forest Service (FS)|
|FUTSE, J - University Of Ghana|
Submitted to: PLOS ONE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/17/2016
Publication Date: 3/29/2016
Citation: Beckley, C.S., Shaban, S., Koku, R., Palmer, G.H., Hudak, A.T., Noh, S.M., Futse, J.E. 2016. Disaggregating tropical disease prevalence by climatic and vegetative zones within tropical West Africa. PLoS One. 11(3):e0152560.
Interpretive Summary: The World Health Organization defines tropical diseases as encompassing “all diseases that occur solely, or principally, in the tropics” and that “in practice, the term is often taken to refer to infectious diseases that thrive in hot, humid conditions”. While tropical infectious disease prevalence is, in general, dependent on a broad number of factors, including economic, demographic, and socio-cultural determinants, rainfall and temperature frequently underlie overall prevalence. This is especially true for arthropod vector-borne diseases for which vector presence, abundance, activity, and seasonality are highly dependent on climate. As a result vector-borne diseases, including “targeted” diseases such as malaria as well as neglected infectious diseases, have a highly skewed distribution with increased prevalence in tropical countries. Specific to Africa, the tendency to incorrectly infer that tropical diseases are uniformly prevalent throughout the roughly 75% of the continent that lies within the tropics has been overcome, at least partially, with solid epidemiologic data, including the data presented here from Ghana. This finer resolution epidemiologic data has at least two important implications. First is that prevalence data can guide treatment, especially in areas where the diagnosis is primarily based on clinical signs. Second is the importance for population immunity. Boundaries where higher prevalence zones, with a correspondingly higher level of population immunity, intersect with zones of lower prevalence and low population immunity create risk for more rapid spread and more severe disease if the underlying transmission determinants change. In this paper we report the testing of the hypothesis of significant differences in pathogen prevalence in Ghanian cattle when comparing vegetation zones, and then determine if weather data and the enhanced vegetation index (EVI) could be used to predict pathogen prevalence. By vegetation zone, tick borne pathogens had a higher prevalence in the coastal savannah as compared to either the Guinea savanna or the semi-deciduous forest, supporting acceptance of the hypothesis and allowing testing of the predictive effect of climatic variables. Weather data and the enhanced vegetation index (EVI) for a three year period for each zone were used in a stepwise multiple linear regression model to identify the statistically significant variables. EVI and rainfall together were most predictive of pathogen prevalence. These findings support the utility of climatic data for understanding vector-borne disease epidemiology on a regional level within the overall context of a tropical country.
Technical Abstract: Tropical infectious disease prevalence is dependent on many socio-cultural determinants. However, rainfall and temperature frequently underlie overall prevalence, particularly for vector-borne diseases. As a result these diseases have increased prevalence in tropical as compared to temperate regions. Specific to tropical Africa, the tendency to incorrectly infer that tropical diseases are uniformly prevalent has been partially overcome with solid epidemiologic data. This finer resolution data is important in multiple contexts, including understanding risk, predictive value in disease diagnosis, and population immunity. We hypothesized that within the context of a tropical climate, vector-borne pathogen prevalence would significantly differ according to zonal differences in rainfall, temperature, relative humidity and vegetation condition. We then determined if these environmental data were predictive of pathogen prevalence. First we determined the prevalence of three major pathogens of cattle, Anaplasma marginale, Babesia bigemina and Theileria spp, in the three vegetation zones where cattle are predominantly raised in Ghana: Guinea savannah, semi-deciduous forest, and coastal savannah. The prevalence of A. marginale was 63%, 26% for Theileria spp and 2% for B. bigemina. A. marginale and Theileria spp. were significantly more prevalent in the coastal savannah as compared to either the Guinea savanna or the semi-deciduous forest, supporting acceptance of the first hypothesis. To test the predictive power of environmental variables, the data over a three year period were considered in best subsets multiple linear regression models predicting prevalence of each pathogen. Corrected Akaike Information Criteria (AICc) were assigned to the alternative models to compare their utility. Competitive models for each response were averaged using AICc weights. Rainfall was most predictive of pathogen prevalence, and EVI also contributed to A. marginale and B. bigemina prevalence. These findings support the utility of environmental data for understanding vector-borne disease epidemiology on a regional level within a tropical environment.