Dramatic decreases of malaria transmission intensities in Ifakara, south-eastern Tanzania since early 2000s

Abstract

Background: Ongoing epidemiological transitions across Africa are particularly evident in fast-growing towns, such as Ifakara in the Kilombero valley, south-eastern Tanzania. This town and its environs (population ~ 70,000) historically experienced moderate to high malaria transmission, mediated mostly by Anopheles gambiae and Anopheles funestus. In early 2000s, malaria transmission [Plasmodium falciparum entomological inoculation rate (PfEIR)] was estimated at ~ 30 infectious bites/person/year (ib/p/yr). This study assessed the PfEIR after 15 years, during which there had been rapid urbanization and expanded use of insecticide-treated nets (ITNs).

Methods: Randomly-selected 110 households were sampled across Ifakara town and four adjacent wards. Mosquitoes were trapped nightly or monthly (June.2015-May.2016) using CDC-light-traps indoors, Suna^® traps outdoors and human landing catches (HLC) indoors and outdoors. All Anopheles mosquitoes were morphologically identified and analysed by ELISA for Plasmodium circumsporozoite proteins. Mosquito blood meals were identified using ELISA, and sub-samples of An. gambiae and An. funestus examined by PCR to distinguish morphologically-similar siblings. Insecticide resistance was assessed using WHO-susceptibility assays, and some Anopheles were dissected to examine ovariole tracheoles for parity.

Results: After 3572 trap-nights, one Plasmodium-infected Anopheles was found (an An. funestus caught outdoors in Katindiuka-ward by HLC), resulting in overall PfEIR of 0.102 ib/p/yr. Nearly 80% of malaria vectors were from Katindiuka and Mlabani wards. Anopheles gambiae densities were higher outdoors (64%) than indoors (36%), but no such difference was observed for An. funestus. All An. funestus and 75% of An. gambiae dissected were parous. Anopheles gambiae complex consisted entirely of Anopheles arabiensis, while An. funestus included 84.2% An. funestus s.s., 4.5% Anopheles rivulorum, 1.4% Anopheles leesoni and 9.9% with unamplified-DNA. Anopheles gambiae were susceptible to bendiocarb and malathion, but resistant to pyrethroids, DDT and pirimiphos-methyl. Most houses had brick walls and/or iron roofs (> 90%), and 52% had screened windows.

Conclusion: Malaria transmission in Ifakara has decreased by > 99% since early-2000s, reaching levels nearly undetectable with current entomological methods. These declines are likely associated with ITNs use, urbanization and improved housing. Remaining risk is now mostly in peri-urban wards, but concerted efforts could further decrease local transmission. Parasitological surveys are required to assess actual prevalence, incidence and importation rates.

Fig. 1

Map of the study areas showing five wards that make up the Ifakara town

Fig. 2

Pictures of sampling methods used: a SUNA trap placed outdoors and b CDC-light trap placed near occupied bed net

Fig. 3

Proportions of indoor and outdoor mosquito catches in the sampling stations for; a An. arabiensis; b An. funestus. This figure illustrates only data collected using human landing catches

Fig. 4

Distribution of mosquitoes in Ifakara town: a hourly distribution of An. funestus; b hourly distribution of An. arabiensis; c monthly distribution of An. funestus and d monthly distribution of An. arabiensis

Fig. 5

Examples of house structures and locations in Ifakara town: a a typical house in Katindiuka ward, showing aquatic breeding habitats of Anopheles mosquitoes (rice fields and water ponds) nearby and b a typical house in the more urban settings of Ifakara town

Fig. 6

Map showing distribution of a An. arabiensis and b An. funestus within the five wards of Ifakara town. All clusters depicting areas with households where the highest densities are most spatially concentrated were first identified, after which statistically significant ones were determined at level of Gi* P value ≤ 0.05, and Gi* Z score ≥ 1.96. The actual Getis-Ord Gi* statistics are provided to illustrate areas with maximum and minimum vector densities