Trypanosomiasis, a disease of humans and animals caused by several species of trypanosomes and spread by tsetse flies is a major constraint to livestock production in 37 countries within the Sub-Saharan region. An estimated 45-50 million cattle are at risk of infection in the region, with an estimated economic loss of up to US$ 1.3 billion in cattle production [1]. Its public health importance has led to attempts to control the disease nationally and regionally with initiatives as Pan Africa tsetse and trypanosomosis eradication program (PATTEC) [2]. These attempts rely on repeated large-scale epidemiological studies and environmental surveys, guiding the design and implementation of intervention strategies. The accuracy of these surveys is limited by use of parasitological diagnostic techniques as microscopy due to low sensitivity [3], and the difficulty in incorporating climatic and environmental data known to influence tsetse distribution, and as a result disease spread [4, 5].

The high costs required to produce tsetse distribution maps through ground-based vector surveys have resulted in few studies looking at the spatial dimension of disease prevalence [6]. However, the use Geographical Information System (GIS) software now makes it cheaper and easier to produce maps which can serve as useful tools for policy discussion, and allow for analysis of factors that would influence the disease patterns [7]. Polymerase Chain Reaction (PCR) diagnostic assays overcome the low sensitivity limitations of parasitological techniques. They are powerful tools for identification and diagnosis of trypanosomes in their hosts and vectors although their high cost and need for elaborate expertise has delayed their adoption [8]. However, PCR assays able to detect all pathogenic trypanosome species in a single reaction have been developed [9, 10]. These reduce the costs of screening a sample from an endemic area by up to five times, and have been suggested as suitable for large-scale epidemiological studies [11].

Results obtained from molecular studies, associated with geo-referenced information concerning vector, cattle distribution and relevant environment parameters, combined in a GIS have the potential of providing more informative study results. This new approach of studying complex pathogenic system is argued to lead to a better evaluation of the risk of infection, allows for effective risk communication, and gives scientific outputs in ways that are understandable to non-scientists [12]. Here we report a study combining use of highly sensitive and species specific PCR assays, with GIS for spatial analysis of infection patterns. The aim is to provide information about the prevalence and distribution AAT in two areas in Western Kenya, and suggest improved control strategies based on the findings.

Attempts to control trypanosomiasis are based on large-scale epidemiological studies, active detection and treatment of confirmed cases, combined with tsetse control programmes [2]. The accuracy of these studies depends on the quality of the study design, and on the sensitivity and specificity of diagnostic methods used. Previous studies have for instance reported T. congolense as the predominant species in cattle in various parts of Africa [13, 14], whereas this study established T. vivax as the most prevalent species. This finding agrees with other studies conducted in Busia district, Kenya [15], and in Tororo, South East Uganda [16]. T. vivax is pathogenic and sometimes causes a fatal hemorrhagic disease.

Overlays with the tsetse maps showed a large proportion of trypanosome cases falling in tsetse-free areas. Assuming the maps are accurate, we suspect that the transmission in those areas is mechanical since most infections are T. vivax, which is transmitted both by tsetse flies and mechanically by biting flies. A longitudinal study comparing bovine trypanosomiasis in tsetse free and tsetse infested areas of northwest Ethiopia found T. vivax to be responsible for 90.9% of the cattle trypanosome cases in the tsetse free zone support this hypothesis [17]. Animals may move between the tsetse-infested and tsetse-free areas although this is unlikely to adequately explain the high infection numbers in tsetse free areas. Farmers in the study region either tether their animals or graze them communally within limited areas.

The presence of mixed infections emphasizes the need for epidemiologists to focus beyond single species to multiple infections with different parasites which are a norm under field conditions [18]. In addition to trypanosomes, animals used in this study may have been infected with other parasites endemic in the region.

Lower trypanosome prevalence in Teso compared with Suba district could be attributed to the fact that unlike Suba, Teso was part of the FITCA program which focused on tsetse control and animal treatment in selected regions [19]. The small sample sizes used in this study limit its power to adequately detect true differences in the two study districts. A similar study using larger sample sizes is recommended. In Teso, a temporal decline of trypanosome prevalence is reported, with up to 40% prevalence in the late 90's, [20] and only 4% in 2005 [15]. Our study reports a much higher prevalence (29%) possibly reflecting higher sensitivity levels attained while using the PCR assays, or a possible lapse in stringent control measures following cessation of the FITCA campaigns. This apparent resurgence raises a question on sustainability of such control programs after donors pull out. Large differences in prevalence as may be obtained by PCR and parasitological methods could have a significant impact on the control strategies selected.

The tsetse maps used here are dated the year 1998, and land usage could have interfered with the tsetse habitat and hence it's distribution. For instance, in Teso district, the status of tsetse may have changed following the FITCA campaigns. Use of impregnated targets and deltamethrin treated cattle has been shown to reduce tsetse populations by 90% and the incidence of bovine trypanosomiasis from over 30% to below 5% in a 6 month period [21].

Whereas this study was not designed to identify the spatial risk factors associated with trypanosomiasis, it shows the importance of simple mapping in helping understand the status of the disease, and possibly guiding future research questions and policy formulation in disease mitigation strategies. For instance, is mechanical transmission of trypanosomiasis significant or is it a case of tsetse habitats having extended to areas captured as tsetse free before? A combination of highly sensitive diagnostic methods which can be extended to study infection in the vector tsetse and GIS provides an opportunity to use evidence based control strategies.

The prevalence maps correlated with tsetse maps show cases lying outside the tsetse areas, mostly being cases of T. vivax which is known to be transmitted both biologically by tsetse flies and mechanically by biting flies. These findings suggest that control of tsetse is not enough and that to clear the possibly mechanically transmitted T. vivax infections, trypanosomes in cattle should also be targeted.