Background
Malaria is the most prevalent mosquito borne disease posing a potential health risk to almost half of the world’s population. The World Health Organization estimated that approximately 216 million people were infected worldwide in 2010 resulting in an estimated 655,000 deaths, of which 91% were in the African region [
1]. In the past few years, reports have indicated a decrease in malaria cases and related deaths in several endemic countries in Africa [
1‐
3]. Although this decline has been attributed to the scale-up of effective malaria control interventions [
1], some reports indicate that in some areas it preceded it [
4]. Of particular interest in this respect is the decline in malaria in north-eastern Tanzania, which has been reported to occur in parallel with a dramatic decline in anopheline mosquitoes in an area with no organized vector control [
5,
6]. The prevalence of lymphatic filariasis, another infection transmitted by anopheline mosquitoes, is also currently decreasing in north-eastern Tanzania. Much of this decrease is due to an ongoing mass drug administration programme [
7], but the decrease in anopheline mosquitoes may also be a contributing factor.
The
Anopheles gambiae s.l. complex serves as an important vector of both malaria and lymphatic filariasis on the East African coast. The complex comprises of seven well recognized sibling species which have only slight morphological differences (not sufficient to make clear distinction between them) but they vary in their ecology and behaviour which is directly reflected in their vectorial capacity [
8‐
11]. Studies conducted on the East African coast have identified
Anopheles gambiae s.s.,
Anopheles arabiensis and
Anopheles merus to be the local sibling species in the
An. gambiae s.l. complex [
12‐
17]. The two fresh water breeding species (
An. gambiae s.s. and
An. arabiensis) differ in their degree of host preference, with the first being strongly anthropophilic [
18‐
20], while the later is more liberal and exhibit zoophilic tendencies especially when alternative mammalian hosts are available [
9,
19]. In areas where domestic animals are kept outside,
An. arabiensis has been reported to have a greater tendency of feeding and resting outdoors, and it has been shown in such areas that human blood index and sporozoite rates are much lower in
An. arabiensis as compared to
An. gambiae s.s. [
21]. Other studies have moreover indicated that
An. gambiae s.s. lives longer than
An. arabiensis[
22,
23]. The salt-water breeding species,
An. merus has been considered a non-vector with exophilic and zoophilic tendencies [
19], but more recent studies have shown that it can be a vector for both malaria and lymphatic filariasis [
12,
24]. Specifically for the coast of north-eastern Tanzania, past research has documented the predominance of
An. gambiae s.s. over the two other members of the complex [
12,
14,
25‐
27].
Due to their diverse ecology and behaviour, identification of individual sibling species of the
An. gambiae s.l. complex is of paramount importance for understanding the epidemiology of the infections they transmit and for setting up the most appropriate control interventions. The observed overall decline in the anopheline population [
6] may have affected the sibling species of the
An. gambiae s.l. complex differently, resulting in change in the sibling species composition and thereby in the transmission characteristics. This study was designed to analyze for a possible change in the relative abundance of members of the
An. gambiae s.l. complex over time (in light of the decline in overall anopheline population) as basis for better understanding the decline in malaria and lymphatic filariasis transmission in north-eastern Tanzania.
Discussion
This study documented the temporal change in relative abundance of sibling species of the
An. gambiae s.l. complex in four villages in north-eastern Tanzania, by comparing the present composition to that seen in previous studies in the same villages and by analysis of archived specimens. Potential confounding associated with differences in sampling methods and in seasonal and spatial variation were taken into consideration when designing the study and analysing the findings. The mosquitoes were collected using CDC light traps, which are effective for indoor collection of host seeking mosquitoes, with a catch that compares well with the standard human landing catch method [
32]. Considering the previously reported decline in anopheline mosquito density [
6], it is noteworthy that the relative abundance of
An. gambiae s.l. complex mosquitoes was considerably higher for the freshly collected mosquitoes in study A (average number per trap per night was 2.74) than what was earlier reported from Kirare [
6] (annual averages of 0.2-0.3
An. gambiae s.l. per trap per night for 2006–2008). This might be due to the fact that mosquitoes for study A were collected during the peak mosquito production season while in Kirare mosquitoes were collected throughout the year. The fact that the abundance also varied considerably between the three villages in study A ( average of 6.2, 1.7 and 0.3 for Tawalani, Vyeru and Kwale villages, respectively) indicates that generalization of the entomological parameters across villages should be done with caution. As the employed light traps mainly collect indoor host-seeking mosquitoes, an attempt was made to assess the outdoor-biting activity by using two outdoor operated MosquitoMagnet™ traps (American Biophysics Corporation, East Greenwich RI) [
33] in Vyeru and Tawalani during the same period. Only eight
An. gambiae s.l. complex mosquitoes were caught in these traps during seven days and nights of trapping, indicating that outdoor biting by the members of the
An. gambiae s.l. complex was at a rather low level in the study villages.
The analysis of the
An. gambiae s.l. complex revealed, as in earlier studies, that
An. gambiae s.s
.,
An. merus and
An. arabiensis were the sibling species found in the study area. Of 585 specimens of the
An. gambiae s.l. complex from study A identified to species level, over three quarters (76.8%) were
An. arabiensis. When comparing the current composition with that reported 2–4 decades ago (Table
2), there had been a marked change in sibling species composition over time. From being by far the most abundant in the past,
An. gambiae s.s
. was now the most rare, whereas
An. arabiensis had changed from being the most rare to the most abundant. The findings on archived mosquitoes in study B confirmed that the process of change in composition of the
An. gambiae s.l. complex also continued during more recent time. Thus, in the earlier survey period (study B1),
An. gambiae s.s and
An. arabiensis were in comparable proportions (39.2% vs 41.9%, respectively). In the later survey period (Study B2) an outstanding shift in composition was noted whereby the proportion of
An. gambiae s.s. had decreased to only 7.7% of what it was in the earlier study period (Study B1) while that of
An. arabiensis had almost doubled.
During the sampling period for study A (rainy season), the population of
An. gambiae s.s. was expected to be at its peak, while that of
An. arabiensis was expected to be low as it normally builds up gradually toward the dry period [
9,
34]. It has moreover earlier been reported [
12] that
An. merus and
An. gambiae s.s. occured in approximately equal proportions during the rainy season where breeding sites for both species were plentiful, but as the fresh water pools dried up the population of the latter diminished while that of the earlier persisted. It therefore appears that the observed shift in sibling species composition of the
An. gambiae s.l. complex was not simply due to seasonal variation, and the reason for the marked change in composition is unclear. Some studies have shown that malaria control interventions such as indoor residual spraying and use of insecticide treated nets while lowering the abundance of anopheline mosquitoes can sometimes also change the relative composition of the
An. gambiae s.l. complex [
35‐
38]. Although no organized vector control activities have been reported in the study areas, it cannot be excluded that even a limited distribution of insecticide treated nets to pregnant women may to some extent have affected the vector complex composition, as reported elsewhere [
39,
40]. However, it appears likely that additional more powerful environmental factors, such as climate change, pollution and/or change in socio-economic standards, may have played an important role.
Examination of the
An. gambiae s.l. for
P. falciparum revealed very few infections (overall rate of 0.24% in the 833 examined specimens), all of which were found in
An. arabiensis. The low infection rate in vector mosquitoes agrees with the reported decline in parasitaemia in human beings [
5]. In contrast,
W. bancrofti infection was found in all the three sibling species, with an overall infection rate of 3.6% (Table
1). It should be noted that these rates are based on all vector-borne stages of the parasite, since the PCR test used cannot distinguish between the different larval stages of
W. bancrofti. Another cause for the high rate of
W. bancrofti in the vectors could be a relatively high prevalence of microfilaraemia in the human population (reported to be 10.6% in Kirare village in October 2008 [
7]) despite the ongoing programme for control of lymphatic filariasis. An earlier study in Tawalani village showed a malaria sporozoite rate of 0.37% and a
W. bancrofti infection rate of 9.5% [
27].
In the course of the PCR identification of members of the
An. gambiae s.l. complex, 39 specimens (mainly from Kirare, see Table
1) could not be identified using the current protocol. However, during detection of infection, 4 of unidentified specimens were positive for
W. bancrofti indicating that probably we were dealing with a filarial vector mosquito that does not belong to the
An. gambiae s.l. complex. Following subsequent inclusion of
Anopheles funestus group primers, as described [
41], 64.1% of the unidentified specimens (including the four
W. bancrofti infected specimens) were found to belong to the
An. funestus group (23
An. funestus s.s., one
Anopheles rivulorum and one
Anopheles leesoni). This emphasize the challenges involved in the identification of light trap collected anopheline specimens, as important morphological diagnostic features (maxillary palps and legs, important for separating
An. gambiae s.l. complex and
An. funestus group) are frequently damaged.
The change in composition of the
An. gambiae s.l. complex obviously has important implications for the transmission of both malaria and lymphatic filariasis. The efficiency of transmission of these infections is closely related to the host choice, feeding habits and longevity of the vectors, which varies between the sibling species [
20,
23,
24,
42‐
44]. While transmission by the more anthropophilic, endophilic and long-living
An. gambiae s.s. is declining, it is taken over by a more adaptable
An. arabiensis.
Anopheles arabiensis is a relatively poor vector compared to
An. gambiae s.s.[
23], and its zoophilic and exophilic behavior moreover renders it less in contact with insecticide treated material (such as impregnated bed nets and insecticide treated walls) thus keeping it at a reduced risk of insecticidal pressure [
37]. The increased role of
An. merus as a potential vector also presents a challenge in the control as this particular mosquito is strongly exophilic, breeds in extensive brackish salt water, and is difficult to control with treated materials and larvicides. Moreover, since the
An. merus population peaks during the dry period (when the salt concentration is optimal) while the population of fresh water species diminishes, this vectorial system will sustain transmission for a much longer period. While these considerations are valid for both malaria and lymphatic filariasis, the effect of the shift in
An. gambiae s.l. sibling species composition on transmission of
W. bancrofti is even more complex, as its transmission in East Africa involves both anophelines (
An. gambiae s.l.,
An. funestus) and culicines (
Culex quinquefasciatus). Although the anophelines in this area are more efficient lymphatic filariasis vectors than the culicines [
7,
19,
45,
46], an observed increasing number of culicines may to some extent compensate for the decreasing number of anophelines. It is therefore likely that the decline in anopheline mosquitoes will have comparatively less impact on the transmission of lymphatic filariasis than that of malaria.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
YAD, MA, KMH, DWM, SMM, EMP and PES conceived the study and participated in its design. YAD coordinated the field and laboratory work, and drafted the manuscript with contributions from PES, MA and DWM. All authors read and approved the final manuscript.