Outdoor attractive targeted sugar bait Phase III trials for malaria control in Kenya, Mali, and Zambia: An individual participant data meta-analysis

This individual participant data meta-analysis of Phase III trials in Kenya, Mali, and Zambia found that while the Westham Sarabi v1.2 ATSB intervention did not significantly reduce malaria incidence or entomological outcomes at the tested deployment levels, a post-hoc analysis revealed a significant 19% reduction in clinical malaria for every 10 bait stations per hectare increase in spatial density, suggesting that higher coverage or optimized dosing strategies may be required for efficacy.

The Gist

Imagine malaria as a relentless army of tiny, invisible soldiers (mosquitoes) trying to invade a village every night to bite people and spread sickness. For years, we've tried to stop them with nets (like building fences) and sprays (like using bug spray). But the enemy is adapting, and we need new weapons.

This paper is a report card on a new weapon called ATSB (Attractive Targeted Sugar Bait). Think of ATSB not as a trap that kills instantly, but as a "poisoned picnic."

The Concept: The Poisoned Picnic

Mosquitoes, just like humans, need sugar for energy. They usually get it from flowers. The ATSB device is a small station hung on the outside of a house. It contains a sweet, sticky treat (date paste) mixed with a tiny amount of poison (dinotefuran).

The idea is simple:

1. Attract: The sweet smell lures the mosquitoes in.
2. Feed: They land and eat the sweet treat.
3. Kill: The poison kills them.

The goal was to set up these "picnics" all over the village to thin out the mosquito army before they could bite anyone.

The Big Experiment

The researchers didn't just test this in one small town. They ran three massive, high-stakes experiments (called Phase III trials) in three different countries: Kenya, Mali, and Zambia.

They treated thousands of houses, hanging two of these "poisoned picnic" stations on the outside wall of every home. They watched for two years to see if fewer children got sick with malaria and if fewer mosquitoes were buzzing around.

The Results: The Picnic Didn't Work (As Expected)

Here is the disappointing news: When they hung two stations per house, the "poisoned picnic" didn't significantly reduce malaria sickness or the number of mosquitoes.

It's as if they set up a few picnic tables in a giant park, but the mosquitoes were too busy eating real flowers elsewhere, or the picnic tables weren't enough to make a dent in the huge mosquito population. The data showed no clear difference between the villages with the stations and the villages without them.

The Twist: It's All About Density

However, the researchers found a fascinating clue while digging deeper into the data. They realized that the number of picnic tables mattered a lot.

• The Analogy: Imagine trying to stop a flood with buckets. If you have one bucket, it won't help. If you have a few, maybe a little. But if you have a wall of buckets, you might stop the water.
• The Finding: The study found that in areas where they managed to hang many more stations per square mile (specifically, 10 extra stations per hectare), malaria cases dropped by about 19%.

But there was a catch: The stations had to be in good condition. If the "picnic tables" were broken, damaged, or missing, the poison didn't work. In some villages, the stations got damaged by weather or people, so the "picnic" wasn't actually happening.

Why Didn't It Work Everywhere?

The paper suggests a few reasons why the standard setup (two stations per house) failed:

1. Not Enough Stations: The villages in Kenya and Zambia were spread out. Hanging two stations on a house wasn't enough to cover the whole neighborhood. The mosquitoes could easily find other places to eat sugar.
2. The Wrong Neighborhood: The math suggests this weapon might work better in cities or crowded towns where houses are packed tight. In a crowded city, you can easily hang 10 stations per block. In a rural village with houses far apart, you'd need to hang an impossible number of stations to get the same effect.
3. Mosquito Preferences: In one specific area of Kenya (near a swamp), the main mosquito was a different type (An. funestus) that seemed to ignore the stations entirely, or the stations might have even made things slightly worse in that specific spot.

The Takeaway

The researchers conclude that while the "poisoned picnic" is a clever idea, the current way of using it (two per house in rural areas) isn't a magic bullet.

• What they learned: To make this work, you might need to hang many more stations in a smaller area, or you need to build a better, tougher station that doesn't break in the rain.
• The Lesson for Future Trials: When testing new tools, scientists need to be very careful about how they pick their test sites. They found that some villages were so different from others (some had way more mosquitoes than others) that it made the results hard to read. Future tests need to be more careful about picking "apples-to-apples" villages to compare.

In short: The "poisoned picnic" didn't save the day in its current form, but the experiment taught us that if we can pack them tighter and keep them in better shape, they might one day become a powerful tool in the fight against malaria.

Technical Summary

Problem Statement
Progress against malaria has stalled, necessitating the identification of new and complementary vector control tools to suppress transmission in high-burden areas. Attractive Targeted Sugar Bait (ATSB) stations represent a new paradigm utilizing mosquitoes' natural sugar-feeding behavior to deliver an oral toxicant (dinotefuran) via an "attract and kill" mechanism. While early proof-of-concept studies in arid settings (Mali and Israel) and modeling suggested potential efficacy, rigorous evidence was required to determine if outdoor deployment of the Westham Sarabi v1.2 ATSB could provide public health benefits when deployed alongside existing standard-of-care interventions (such as insecticide-treated nets). The specific challenge addressed was evaluating the efficacy of this new tool class in diverse, high-transmission rural settings (Kenya, Mali, and Zambia) where vector species and ecological contexts vary.

Methodology
This study presents an individual participant data (IPD) meta-analysis of three large-scale, Phase III, cluster-randomized controlled trials (cRCTs) conducted between 2021 and 2024 in Kenya, Mali, and Zambia.

• Intervention: The Westham Sarabi v1.2 ATSB stations (containing 0.11% dinotefuran) were deployed on exterior walls of residential structures at a rate of two per structure.
• Design: The trials followed a master protocol but were conducted independently. The unit of randomization was the village (cluster).
• Outcomes: The primary epidemiological outcome was clinical malaria incidence in children (ages 1–14 in Kenya/Zambia; 5–14 in Mali). The primary entomological outcome was the parity of dominant vector species. Secondary outcomes included Plasmodium falciparum infection prevalence, vector abundance, landing rates, and sporozoite positivity.
• Analysis: The authors conducted both pooled meta-analyses using site-level effect estimates and IPD meta-analyses using a two-stage approach. Models accounted for country as a fixed effect and cluster as a random effect.
• Post-hoc Analyses: The study included dose-response analyses to assess the relationship between ATSB spatial density (adjusted for coverage and physical condition) and clinical malaria incidence. Subgroup analyses were performed to explore effect modification by predominant vector species (An. gambiae s.l. vs. An. funestus s.l.).

Key Results

• Primary Efficacy: The meta-analysis found no statistically significant differences between intervention and control arms for any of the primary or secondary epidemiological or entomological outcomes.
  • Clinical malaria incidence: IRR 0.91 (95% CI 0.82–1.02) in pooled analysis; IRR 0.97 (95% CI 0.84–1.12) in IPD analysis.
  • P. falciparum prevalence: OR 0.93 (95% CI 0.80–1.10) in pooled analysis.
  • Entomological indicators (parity, abundance, landing rate, sporozoite positivity) showed no statistically significant changes, though some point estimates suggested slight increases in parity and sporozoite positivity in the intervention arm.
• Dose-Response Relationship: A significant dose-response relationship was identified. For every 10-unit increase in coverage-adjusted cluster-level ATSB density (specifically counting devices in good condition), clinical malaria incidence decreased by 19% (IRR 0.81, 95% CI 0.74–0.89, p<0.001). This effect was contingent on the physical condition of the devices; unadjusted density measures showed no association.
• Vector Species Interaction: In Kenya, where vector species could be classified across all clusters, an interaction was observed. In clusters where An. funestus was predominant, clinical malaria incidence was higher in the intervention arm (IRR 1.71), whereas in An. gambiae predominant clusters, the incidence was lower (IRR 0.76), though the latter was not statistically significant.
• Heterogeneity: The study noted substantial heterogeneity between clusters, particularly in Kenya, where the coefficient of variation (CV) for clinical incidence was higher than anticipated (0.75), complicating the power to detect effects.

Significance and Claims
The paper concludes that the deployment of two Westham Sarabi v1.2 ATSB stations per residential structure, as tested in these rural settings, did not result in statistically significant reductions in clinical malaria incidence or entomological indicators. However, the identification of a dose-response relationship suggests that the intervention may be efficacious under specific deployment conditions, particularly where higher spatial densities of functional devices can be achieved. The authors suggest that such high densities may be more feasible in peri-urban or urban settings with higher housing density, rather than the rural settings tested.

The study highlights critical lessons for future vector control cRCTs, emphasizing the need for:

1. Comprehensive baseline data collection to identify outliers and balance vector bionomics.
2. Collecting a limited but balanced set of entomological outcomes across all trial clusters to ensure adequate power.
3. Considering sequential rather than parallel trial designs for new interventions to allow for protocol refinement based on interim findings.
4. Re-evaluating sample size assumptions based on observed coefficients of variation, as high heterogeneity can render trials infeasible.

Ultimately, the authors posit that while the current deployment approach did not demonstrate efficacy in these specific rural contexts, the dose-response findings warrant further research into alternative deployment strategies, optimal spatial densities, and product refinements to enhance durability and attractiveness.