Mosquitoes, men and malaria in Kenya. A study on ecological factors affecting malaria risk

Despite various control efforts, malaria remains a major cause of illness and death on the African continent. In the near future, there may even be an increased risk of malaria as a result of changes in our natural environment, such as global warming, deforestation and urbanisation. There is an urgent need to assess the impact of these changes on malaria vector populations and malaria risk. The aim of this thesis is to elucidate mechanisms that regulate the population dynamics of malaria vectors and relate these to malaria transmission and malaria risk. Field and laboratory studies were carried out in westernKenya, where malaria seriously affects the daily life of humans. With this information, local malaria risk models will be developed that may be used to design tailor-made approaches for malaria prevention and control.Many physical, biological and socio-economic factors affect the risk of malaria. Temperature, rainfall and humidity determine mosquito survival and development, whereas the presence of mosquito breeding sites, vegetation, land-use, house construction and the use of preventive measures mainly determine vector-host contact (Chapter 2). Many of these factors are conduciveformalaria transmission in western Kenya. The local population, therefore, spends much money on malaria preventive measures and health care (Chapter 3). These costs could be averted if a good and affordable health care system is in place.Within the aquatic habitats of An. gambiae sensu lato , the main vector of malaria in sub-SaharanAfrica, larvae may be observed feeding and interacting with each other. These larvae are under a high pressure to develop to the adult stage as rapidly as possible, since their habitat may dry up within a few days if no rain falls. Nevertheless, eggs and larvae may survive on damp soil for a few days (Chapter 5). Moreover, first-instar larvae that hatch from eggs on damp soil may reach nearby water by active dispersal (Chapter 5). High densities may lead to opportunities for cannibalism and predation among larvae of An. gambiae s.l. (Chapter 6). This is mostly affected by the amount of space available, but not by the amount of food (Chapter 7). The fact that larvae aggregate (Chapter 10) or experience a decreasing habitat size if no rain falls, will probably lead to frequent interactions between larvae and thus to cannibalism and predation in the field. Using DNA-fingerprinting techniques (microsatellite markers), we found that female An. arabiensis mosquitoes may distribute their eggs over at least two sites (Chapter 8). This may be advantageous from an evolutionary perspective, since breeding sites can dry up rapidly. Our results on egg and larval survival in desiccating habitats, cannibalism and predation, and oviposition strategy suggested that these mechanisms are important determinants of vector population dynamics.African highlands are considered to be most at risk of malaria, since these areas are currently free of malaria. The local population will not have developed a protective immunity. Our studies revealed that adult vectors survive in a highland area, but the larval stages do not develop and die before pupation as a result of the cool temperatures in this area (Chapter 4). Some breeding sites of malaria vectors were found in the area (Chapter 11), but they probably contribute little to the adult vector population. Therefore, the malaria infections observed among school children in this area (Chapter 9) are probably the result of mosquitoes that disperse from lower to higher elevations, survive inside houses and infect people locally. The results suggest that a temperature increase of a few degrees may lead to outbreaks of malaria once vector populations get established in highland areas. At lower elevations, but close to the highland area, malaria risk was strikingly different with malaria being transmitted year round. Interestingly, significant temporal variation in malaria prevalence and morbidity between years was observed, probably as a result of an unusual dry period during the first study year (Chapter 9). We identified rainfall and evapotranspiration as important drivers of vector abundance and species composition (Chapter 10). The estimated number of infective bites (Entomological Inoculation Rate or EIR) could not explain malaria prevalence one or two months later, but a lower EIR during the first (dry) year of study was associated with a lower malaria prevalence and a higher morbidity.The main conclusion from this thesis is that ecological factors may cause large differences in malaria transmission and malaria risk on a relatively small spatial scale. Changes in these factors may result in epidemics in highland areas. Additionally, they may alter malaria risk in lowland areas, but further studies are required to investigate the role of immunity development in the human population. Ultimately, information gained from these studies should be integrated with remotely sensed data to develop accurate models for malaria risk assessment on a wider spatial scale.