Background
Urban malaria is an emerging health problem in Africa [
1]. Some reports indicate that the rapid demographic growth of some urban areas presents a serious challenge for the control of endemic diseases such as malaria [
1,
2]. Several factors influence the epidemiology and severity of malaria disease in urban areas; these factors include the geographical characteristics of the town, the level of immunity in the local population, the extent of unplanned urbanization, urban agriculture development, and the degree of human migration from rural to urban areas [
2,
3]. Although unplanned urbanization characterized by poor housing, lack of sanitation and inadequate surface water drainage is recognized as one of the major factors maintaining the presence of
Anopheles gambiae in some urban settings [
2], it has not been extensively studied in Cameroon.
In Cameroon, malaria affects over 90% of the population and is responsible for 35% of the annual mortality [
4]. Despite the regular occurrence of malaria cases in urban settings, only a limited number of studies have addressed the epidemiology of the disease in such settings. Entomological studies in the city of Yaoundé, reported low malaria transmission rates in central districts, but high transmission rates in peripheral locations [
5‐
7]. Parasitological studies on the other hand, revealed an annual prevalence of 35% of
Plasmodium falciparum asexual parasites carriers, a seasonal variation of parasites prevalence in the population and, a potential infectious reservoir dominated by the age group ranging from 0 – 15 years [
8]. The city of Douala, the largest in Cameroon, has, paradoxically, received less attention. Although there are some similarities between Douala and Yaoundé, several climatic and environmental contrasts exist. For example, Douala is a port city close to the Atlantic Ocean with year-round rainfall and a high level of spontaneous urbanization compared with Yaoundé, which is situated inland at high altitude. All this making that Yaoundé and Douala could experience distinctive malaria epidemiological conditions. Since the 1950s, successive changes have been made to the Cameroon National Policy for Malaria Control. From the 1950s to the 1960s malaria control efforts were mainly oriented on indoors residual spraying of insecticides, in the 1980s primary health care services were promoted, and in the 1990s insecticide-treated bed nets came into use. A lack of knowledge about malaria epidemiology in different contexts has undoubtedly contributed to variations in the success of such measures to combat the disease [
9]. Thus, better understanding of the factors affecting malaria epidemiology in Douala could assist efforts towards malaria control in the city.
This work falls into a series of studies aimed at assessing the influence of urbanization on malaria vector bionomic and malaria transmission in the two major urban cities of Cameroon: Douala and Yaoundé. A recent study conducted in the city of Douala, revealed the presence of a number of mosquito breeding sites that persisted all year long as well as a high prevalence of mosquito resistance to DDT and permethrin insecticides [
10]. Here, we present the results of a 12-month study on the dynamics of malaria transmission and mosquito burden in the district Ndogpassi Douala, which contains hundreds of thousands of urban dwellers.
Discussion
Our data indicate that frequent malaria transmission occurs in Douala city, a finding that is consistent with the majority of studies conducted in urban settings within sub-Saharan Africa [
20‐
25]. The mosquito burden was primarily Culex, followed by
A. gambiae, and Mansonia. The large number of Culex suggests that the area is experiencing difficulties with water and waste management of the type that is common to unplanned urbanized areas in Africa [
26,
27]. The
An. gambiae densities were far greater than what might have been expected, thus supporting the possibility of adaptation of this species to urban areas [
28]. Although the
An. gambiae biting densities were subject to seasonal fluctuations, this species was present all year long owing to the presence of abundant collections of standing water in the district. However,
An. gambiae samples consisted almost exclusively of the M form, a finding consistent with earlier studies conducted in Cameroon that indicated the high prevalence of this molecular form in coastal areas [
29‐
31]. The predominance of the M form over the S form in the study area is most likely associated with its higher tolerance to salinity (Costantini, unpublished data).
Human landing catches appeared to be far more efficient than CDC light traps. This observation is consistent with data from several studies comparing the efficacy of HLC with a diverse range of mosquito collection methods [
32,
33]. Indeed during HLC, mosquitoes are attracted by both visual and chemical stimuli, whereas CDC light traps use only visual stimuli. However, the fact that HLC was undertaken outdoors while CDC LT was conducted indoors could have overestimated the efficiency of HLC compared with CDC LT. Because of the personal and property related security issues in Douala, household owners did not permit us to conduct indoor HLC; this prompted us to adjust the study design, whereby HLC was conducted exclusively outdoors while CDC LT continued to be placed indoors. The number of species collected fluctuated substantially between the two methods but was in agreement with previous studies that have indicated that the efficiency of a collection method can vary according to the composition of the mosquito species present, mosquito densities, availability of alternative hosts, and city lighting [
33‐
35]. The sampling efficiency of both collection methods was found to be independent of vector abundance. Although comparisons between collection methods commonly use log transformation of direct counts, the method used here would have little bias because it uses log(x + 1) instead of log(x) [
19]. However, use of negative binomial regression analysis or Bayesian estimations might be more appropriate for some studies [
19].
Malaria transmission estimates in Ndogpassi district were relatively high compared to records from similar urban settings [
27,
36,
37]. The data are consistent with parasitological and clinical records revealing a high prevalence of malarial disease in the population, however [
38]. Malaria transmission was detected only during the rainy season, despite the continuous presence of
An. gambiae and could probably be associated to the low infectiousness of the human reservoir [
8,
21,
36,
39] alternatively, it could reflect the influence of high temperature and low humidity on the longevity and vectorial capacity of
An. gambiae during the dry season. Several factors can influence parasite prevalence in humans and mosquito infectivity; these include, for example, the use of protective measures against mosquitoes, health seeking behaviors, socio-economic status, and immunity [
40‐
42].
Although malaria transmission estimates were not assessed across the city, an earlier study suggested that there might be a heterogeneous risk for malaria transmission that is dependent on the distribution of
An. gambiae[
10]. The situation in Douala could be similar to that of Yaoundé where malaria transmission has been estimated at 277–365 infectious bites per human per year in the peripheral districts, and 3–33 infectious bites per human per year in the central districts [
7,
8,
27]. According to Robert et al. [
37], the annual inoculation rates in sub-Saharan Africa could be as high as 7.1 in the city centers, 45.8 in peri-urban areas and 167.7 in rural areas. The data from this study exceeded the anticipated estimates for central districts and could indicate that the high malaria prevalence is maintained by frequent migration of populations from highly endemic rural areas to the city. Moreover, the fact that no transmission was detected during certain periods of the year does not totally exclude the possibility of perennial malaria transmission since clinical cases were reported all year long in the local health care center (data not shown). Clearly, further studies for are needed to obtain better understanding of all the factors influencing malaria transmission in Douala.
The level of
An. gambiae susceptibility to DDT and pyrethroid insecticides, and the frequency of the
kdr allele in the
An. gambiae population were in agreement with previous studies [
10,
43]. The high prevalence of insecticide resistance in the mosquito population of Ndogpassi district can be attributed to the frequent use of chemical insecticides for market gardening and personal protection [
44]. Apart from insecticide-treated bed nets, insecticide sprays and coils were reported to be in regular use in households [
44], whereas up to eight different insecticide mixtures including pyrethroids, organochlorines, organophosphorus and carbamates have also been reported to be in regular use for urban farming in Cameroon [
45]. A direct association between urban farming and increased insecticide resistance in malaria vectors has been reported in several studies [
23,
25,
45,
46]. Despite the high prevalence of the
kdr allele in mosquito populations, and its close association with resistance to both permethrin and DDT [
45], this resistance mechanism might not be fully responsible for all of the insecticide resistance phenotypes that exist. Further studies are underway to explore the importance of metabolic resistance to insecticides in mosquito populations. The rapid expansion of insecticide resistance is becoming a serious challenge for malaria control across Africa that calls for sustainable solutions [
47,
48].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CAN, PAA conceived of and designed the study protocol. BDT, RTF, PK, BTF, CN, PAA, and CAN participated in the field sample collections. BDT, BTF, CN, and CAN carried out the molecular and ELISA analyses; LGL, TT, and PK helped to draft and revise the manuscript. CAN was responsible for data analysis and interpretation and also wrote the paper. All authors have read and approved the final version.