Trial location
The Zambia ATSB phase III trial site was located in three adjacent districts in Western Province: Kaoma, Luampa, and Nkeyema districts (Fig. 1). Western Province is approximately 400 kilometers from the capital of Zambia, Lusaka. Baseline data was collected in 85 rural clusters of which 70 were included in the trial. The 15 clusters included in the baseline and excluded from the main trial were dropped due to accessibility challenges, high refusals, or very low malaria prevalence. The 70 trial clusters span a geographic distance of approximately 175 kilometers. The clusters were created using a K means algorithm and satellite imagery to draw areas with a minimum of 250 households to meet sample size requirements (Ashton et al., in preparation).
Environment
Climate
Western Zambia has a tropical climate with an average annual rainfall of approximately 1000 mm; 93% of the total rainfall occurs between November and March (5, 6). The total rainfall from November 2020 to March 2021 was 1139mm (Kaoma Meteorological Station, unpublished data). The average daily maximum temperature during this period was 30.3 degrees Celsius and the average daily minimum was 18.9 degrees Celsius. A cool dry season typically occurs from May through August followed by a hot dry period from September through October. Total rainfall from May to October 2021 during the dry season was 59.6mm, maximum daily temperature was 30.7 degrees Celsius, and minimum daily temperature was 11.6 degrees Celsius (Kaoma Meteorological Station, unpublished data).
Land cover and use
The dominant soils in Western Province are Kalahari sands which favor tree growth (6). The study site districts of Kaoma, Nkeyema, and Luampa are predominantly Central Zambezian Miombo Woodlands which are characterized by Miombo trees (Brachystegia genus). There were 339 different nectar producing plant species observed across the trial site (Nicholas Wightman, unpublished report). In particular, the Miombo woodlands are rich in edible indigenous fruit trees (7). These indigenous tree species are found in the larger landscape between households. Mango trees (Mangifera indica) which are planted for fruit and shade, are the most common household tree (Nicholas Wightman, unpublished report). The site also includes seasonally flooded grasslands and dry forests (6). The two main water bodies in the trial site are the Luampa and Luena Rivers. During the rainy season, flooding is common in areas surrounding the rivers, with pools forming as the flooding recedes.
Several trees and plant species found in the trial site are potential sources of sugar meals for malaria vectors. Trees appear to be the main source of sugar production. Mangifera indica (mango trees) and Senna siamea (evergreen trees) are predominant in the trial site and are known to be attractive to An. gambiae (8, Nicholas Wightman, unpublished report). Herbaceous plants that are known to be attractive to mosquitoes and are also common in the study site include Bidens pilosa (lackjack), Senna occidentalis (coffeeweed), Manihot sculenta (cassava) and Ipomoea batatas (sweet potato) (Nicholas Wightman, unpublished report).
Households in this region rely primarily on commercial and subsistence farming. During the trial baseline study, 88% of households across the trial site reported farming or gardening as their primary occupation. The main crops grown were maize, cassava, rice, millet, and vegetables (9, 10). A subset of the population also reared cattle. Other reported livelihoods include tobacco production in Nkeyema district and exploitation of trees for charcoal production (Nicholas Wightman, unpublished report). Households typically have agricultural plots near their living structure(s) to cultivate maize and cassava ranging in size from one to two hectares (see Fig. 3) (Nicholas Wightman, unpublished report). Areas directly surrounding living structures are typically swept clean with little to no vegetation other than trees providing shade and/or fruit (see Fig. 2).
Photo credit: Nicholas Wightman
Population
The combined population of Kaoma, Nkeyema, and Luampa districts was estimated to be 246,785 during the 2022 national census (11). A baseline enumeration of the study area estimated a population of 122,023 in the 70 trial clusters. Households were defined as a person or group of persons, related or unrelated, who live together in the same dwelling compound under one household head and share a common source of food. The average household size in Western Province was 4.7 people (11). All trial clusters were rural. These clusters covered a range of geographic sizes and housing densities. In general, the trial clusters were sparsely built with ~ 0.25 structures per hectare (median 0.36 IQR (0.19–1.24) across clusters). The baseline trial survey found that houses were primarily constructed with thatch/leaf roofs (72%), walls with bamboo/wood and mud (46%), and earth/sand floors (86%). Some houses had corrugated iron roofs (28%) and walls made of stone with mud (30%) (see Fig. 4). For all house types, 21.1% had closed eaves and windows that are closed/sealed leaving many houses not fully sealed and with potential entry points for mosquitos.
Photo credit: Erica Orange
Health system
There were three levels of health facilities available across the trial site including health posts (subdistrict level), health centers (district level), and hospitals (district level 1) (12). There was a total of 29 health facilities across the trial site, of which the majority were health posts. Health posts are staffed by one or more community health assistants that typically have received a standardized twelve-month training on disease screening and prevention (13). During the baseline study, 97% of people within the study area reported seeking care for fever from a health facility inclusive of hospital, health center, or health post.
Table 1
Overview of health facility type in the trial site
Facility type | Description |
District hospital | • Serve a population of 80,000 to 200,000 people • District health officers provide services |
Health center | • Serve a population of 10,000 people • Staffed by a clinical officer, nurse, or environmental health technician |
Health post | • Serve a population of 3,500 people in rural areas • Within 5km radius for sparsely populated area • Staffed by community health assistants that also provide community-level services. |
Source: The National Community Health Strategy: 2019–2021 (13)
COVID-19
Zambia experienced four waves of COVID-19: July 2020, January to February 2021, June to July 2021, and December 2021 to January 2022 (14). Following the fourth wave, Western Zambia recorded 18,887 cumulative cases and 129 deaths (15). COVID-19 vaccines were introduced in Zambia in April 2021 with Johnson and Johnson and AstraZeneca being the predominant vaccines in the trial site.
Malaria
Malaria vectors
Overall Anopheles spp. diversity in the study site was high. During pre-trial entomological feeding studies from March to May 2021 in a subset of 10 clusters within Kaoma and Nkeyema districts, 14 different anopheles’ species were morphologically identified from overnight indoor and outdoor CDC ultraviolet light trap collections during 7,600 nightly collection sites (3). Of these, An. funestus s.l., An. squamosus, An. tchekedii, and An. coustani were the most abundant, with An. gambiae s.l. present in low numbers (Table 2).
Table 2
The mean number of Anopheles mosquitoes collected in US CDC ultraviolet light traps (CDC UV LT), per household collection night, from March to May, 2021 across the Zambia ATSB trial site.
Species | Mean number of males | Mean number of females | Mean total number | % of total N = 20.08 |
An. funestus | 0.27 | 5.99 | 6.26 | 31% |
An. squamosus | 0.02 | 4.02 | 4.03 | 20% |
An. tchekedii | 0.01 | 3.80 | 3.81 | 19% |
An. coustani | 0.01 | 3.76 | 3.77 | 19% |
An. maculipalpis | 0.00 | 0.49 | 0.49 | 2% |
An. gambaie s.l. | 0.01 | 0.45 | 0.45 | 2% |
An. brunnipes | 0.00 | 0.12 | 0.12 | 1% |
An. rufipes | 0.00 | 0.11 | 0.11 | 1% |
An. gibbinsi | 0.00 | 0.08 | 0.08 | < 1% |
An. pretoriensis | 0.00 | 0.01 | 0.01 | < 1% |
An. implexus | 0.00 | 0.00 | 0.00 | < 1% |
An. pharoensis | 0.00 | 0.00 | 0.00 | < 1% |
Other Anopheles | 0.01 | 0.72 | 0.72 | 4% |
Not identified | 0.00 | 0.21 | 0.22 | 1% |
Total | 0.33 | 19.75 | 20.08 | 100% |
During this study, three groups of mosquitoes tested positive for Plasmodium falciparum (Pf) sporozoites assessed by standard Pf ELISA screening: An. funestus s.l. (sporozoite rate [SR] = 3.2%), An. gambiae s.l. (SR = 0.2%), and An. squamosus (SR = 0.2%) (3). An. funestus was both the most abundant Anopheles spp. mosquito collected as well as the dominant malaria vector, representing 95% of all infectious mosquitoes collected in 2021. Infectious An. funestus mosquitoes were equally likely to be collected in light traps set up indoors or outdoors (3).
A subsample of mosquitoes was screened for the presence of natural sugar meals using a cold anthrone test (16, 17). The cold anthrone test identified mosquitoes that have recently (within several hours) acquired a natural sugar meal (16, 17). Results suggested variable natural sugar feeding rates by species (Table 3). Though nearly half (47%) of all An. funestus screened were positive for a recent sugar meal, this proportion varied by cluster and by month, ranging from 19–67% (data not shown).
Table 3
The proportion of mosquitoes positive for recent sugar meals from March to May, 2021 across the Zambia ASB trial site.
Species | Number tested | Proportion Positive |
An. tchekedii | 1,826 | 21% |
An. funestus | 1,483 | 47% |
An. coustani | 1,283 | 13% |
An. squamousus | 1,037 | 10% |
An. pharoensis | 673 | 5% |
An. gibbinsi | 619 | 20% |
An. brunnipes | 84 | 25% |
An. pretoriensis | 61 | 28% |
An. gambiae | 30 | 23% |
An. rufipes | 19 | 53% |
An. implexus | 17 | 6% |
Total | 7,132 | 22%ve |
Finally, standard WHO tube bioassay tests (18) indicated high levels of pyrethroid resistance in local An. funestus (24-hour mortality ranging from 46–60%) and An. gambiae (24-hour mortality ranging from 67–80%) from the study site. Insecticide resistance has not yet been detected for the primary chemical used for IRS in the trial area, clothianidin, nor for other IRS chemicals including pirimiphos methyl. Resistance has also not been detected for the active ingredient used in the Sarabi 1.2 ATSB, dinotefuran (3).
Malaria burden
Malaria was endemic in the trial site and infections peaked following the annual rainy season (19). In Western Province, the estimated malaria incidence was 785 per 1,000 people in 2021 (19). The trial baseline study found P.f. prevalence by rapid diagnostic test (RDT) was 50% among people ages six months and older during peak transmission season.
Other malaria vector control interventions in the trial site
In the trial site, per the National Malaria Elimination Strategic Plan, universal access to vector control interventions was defined as households having access to one ITN per two people or IRS within the past 12 months (20). Three IRS and ITN campaigns were conducted in the trial site using a mosaic approach between November 2020 and November 2022 (see Fig. 5 below). Before the start of the rainy season, the National Malaria Elimination Centre (NMEC) led annual IRS campaigns using Fludora(R) Fusion WP-SB 56.25 (clothianidin and deltamethrin) in a subset of health facility catchment areas that were selected during an annual microplanning exercise. IRS was conducted in 42 of the clusters in 2020, 42 clusters in 2021, and 21 clusters in 2022. In 2020, the NMEC-led ITN campaign directly followed the IRS campaign and included distribution of Permanet (R) 2.0 (deltamethrin) ITNs. Two supplemental ITN distributions were implemented by the trial team within the study clusters. The first was in February 2022 and was implemented to address high demand for ITNs in the context of trial activities. A total of 28,908 Permanet (R) 2.0 (deltamethrin) ITNs were distributed with a strategy of distributing one ITN per household across the trial site. To address potential gaps in ITN coverage following the 2020 NMEC-campaign, a total of 59,051 VEERALIN(R)LN (alpha cypermethrin plus pyrethroid piperonyl butoxide [PBO]) ITNs were distributed in September 2022 to the 48 clusters that were not fully covered by NMEC-led IRS. During this distribution, one ITN was provided for every two residents in the household per the national strategy. As a result of these efforts, the trial site had high levels of coverage with vector control interventions with more than 60% of households having access to at least 1 ITN per 2 people or IRS.
Case management in the trial site
Across the trial site, about half (47%) of people of all ages with fever in the past two weeks reported seeking care from a formal provider during the baseline survey. Approximately 80% of people with a fever who sought treatment from a formal health provider reported being tested for malaria. In this setting, care for suspected malaria was sought almost exclusively from public/government health facility providers. A common limiting factor to the effectiveness of malaria case management is stockouts of key commodities including malaria RDTs and treatments. During the ATSB trial, the study team noted periods of commodity stockouts reported by study participants which may have contributed to suboptimal case management of suspected malaria across the trial site.