Spatial distribution of malaria transmission in relationship to "Anopheles gambiae" complex members in Sudan savanna and irrigated rice cultivation areas of Mali

Malaria remains a major public health problem that is exacerbated by poor
implementation of control measures, and by the spread of drug-resistant parasites and
insecticide resistant vectors. Preventive measures, including those targeted at vectors, are one
of the four basic elements of the global malaria control strategy. The control methods to use
should be selective and specific to the control area. The success of the approach of selective
and targeted interventions requires a good stratification of control areas, which should be
based on mapping of malaria risk and vector species distribution.
The goal of this thesis was to enhance our understanding of the relationship between
the distribution of members of Anopheles gambiae complex and climatic and environmental
conditions, to describe their spatial and temporal distribution, to quantify their unique
contribution to malaria transmission, and to produce attributed malaria risk maps of Mali. We
used Bayesian geostatistical modeling, implemented via Markov chain Monte Carlo
simulation (MCMC), which can quantify the relationship between environmental factors and
the species distribution by taking into account the spatial dependence present in the data in a
flexible way that allows simultaneous estimation of all model parameters. In addition,
Bayesian kriging enables model-based prediction together with the prediction error, a feature which is not possible in the classical kriging.
The analyses described in chapters 2 and 3 identified environmental factors related to
the distribution of a) the two major species (An. arabiensis and An. gambiae s.s.) which
compose the An. gambiae complex and b) the chromosomal (Bamako, Mopti, Savanna
Hybrids) forms of An. gambiae s.s., and produced maps of the geographical distribution of the
species and chromosomal forms. Estimation of the contribution of species and chromosomal
forms to malaria transmission in Mali is described in Chapter 4; the spatio-temporal
distribution of An. gambiae complex densities and its chromosomal (Mopti, Bamako, Savanna, Hybrids) forms in a Sudan savanna village is examined in Chapter 5; the
investigation of malaria vector ecology during the dry season and its implication for vector
control is described in Chapter 6, and Chapter 7 presents the spatial pattern of malaria
transmission in the rice cultivation area of the Office du Niger.
The maps produced in chapters 2 & 3 showed higher frequencies of An. arabiensis in
the drier Savanna areas and An. gambiae s.s. in the flooded/irrigated areas of the inner delta of
Niger river, the southern Savanna, along rivers and in the Sahel. The Mopti form was found in
the same ecological area as An. arabiensis. In addition, it occupied the flooded/irrigated areas
of the inner delta of Niger River. The Savanna form prefers the Sudan Savanna areas and the
Bamako form was confined around Bamako city and in part of Sikasso region (South of
Mali). Analyses in Chapter 4 indicated that high malaria risk was associated with insecticide
resistance gene (kdr) carriers (Bamako/Savanna chromosomal) and Hybrids compared to the
non-carriers An. arabiensis and the Mopti chromosomal form, although the association was
not significant. The attributed risk maps of the different species and subspecies indicated that
in the middle West and South East part of the country malaria transmission risk is mainly due
to An. arabiensis, in the irrigated/flooded areas malaria risk is attributed to the Mopti form, in the southern part to the Savanna/Bamako forms and in the southern areas of the region of
Kayes to the hybrids. Thus these results suggest that insecticide control measures must be
strengthened in the Sahelian (epidemic prone area) and irrigated/flooded areas where An.
arabiensis and the Mopti chromosomal form, which have no or lower frequency of insecticide
resistance gene, prevail. Any vector control by means of insecticides in the Southern part of
the country, where the S molecular form (Savanna and Bamako) predominates, must be
accompanied by a close insecticide resistance monitoring system. The analyses carried out in Chapter 5 and 6 on the spatial distribution of the sibling
species of An. gambiae complex in a savanna village showed that the distribution of mosquito
densities was concentric with higher densities clustering at the periphery of the village at the
beginning of the rainy season and during the dry season. This distribution was patchy during
the middle and the end of the rainy season. The chromosomal forms were sympatric
throughout the seasons. There was a spatial clustering in their relative frequency distribution
changing over time in the village. The Mopti chromosomal form was the most abundant at the
beginning and middle of the rainy season and the Bamako form at the end of the rainy season.
Larval habitats monitoring showed that in the main village of Bancoumana nearly all larval
habitats were human-made, rain-dependent and dried out 10-12 weeks after the end of the
rainy season. At the same time, numerous natural puddles highly productive for anopheline
larvae even during the dry season were located in the fishermen’s hamlets. These were
adjacent to the receding Niger River bed and 5 km away from the main village. Larval
habitats in Bancoumana were re-colonized shortly after rainfall suggesting that mosquitoes
emerging from the riverbed are an important source for the rain-fed water bodies of
Bancoumana. This observation indicates that control interventions targeting the Mopti form
should be implemented at the beginning and middle of the rainy season, while those targeting the Bamako form should be done at the end of the rainy season. In addition, appropriate
vector control implemented in the fishermen’s hamlet during the dry season and at the
periphery of the main village at the beginning of the rainy season may be feasible, sustainable
at low cost and may ameliorate malaria transmission in this area.
In chapter 7, the analyses of malaria transmission parameters in the rice cultivation
area of the Office du Niger indicated a strong spatial correlation in mosquito densities, which
is related to the rice cultivation environment. However, the spatial correlation observed in the
parous rate (PR) and human blood index (HBI) was weak suggesting that these parameters are more closely related to local conditions such as population behavior and economic status,
and/or the presence of animals rather than similar environment over large areas. Since both
the PR and HBI measure the vector-human contact rate, and hence the potential for malaria
transmission intensity, attention must be paid to the local variations when implementing
control strategies in rice cultivation areas.
This work makes a substantial contribution to the mapping of the spatial distribution
of malaria vector species and subspecies which was previously limited by the lack of field
data and appropriate statistical analyses. It also provides valuable information for
conventional vector control as well as future implementation for genetically manipulated
mosquitoes control method.