Larviciding shows variable malaria control impact across African villages, complicating scale-up

Malaria control across Africa relies heavily on insecticide-treated bed nets and prompt treatment of clinical cases, yet these tools alone cannot push the disease toward elimination. The protection they offer erodes over time—nets lose potency, develop holes, get discarded—leaving gaps that widen across the three-year cycle between mass distribution campaigns. Researchers have begun testing an older strategy: larval source management, or LSM, which targets mosquito breeding sites before adult vectors emerge. The question is whether this approach, abandoned decades ago in favor of insecticide spraying, can meaningfully reduce malaria transmission when deployed alongside modern interventions.

Two trials in contrasting settings offer the first rigorous evidence. Between 2004 and 2006, researchers in Western Kenya worked in six highland valley communities at elevations between 1,453 and 1,632 meters, where recent deforestation and swamp clearance had left exposed water bodies ideal for mosquito breeding. Transmission there was year-round, tracking seasonal rains. A second trial ran from 2019 to 2020 in Côte d'Ivoire, a region receiving 1,200 to 1,400 millimeters of annual rainfall with a distinct six-month dry season. Both trials applied Bacillus thuringiensis israelensis, or Bti, a larvicide that kills mosquito larvae in water. In Kenya, larviciding began as insecticide-treated nets were first being introduced to communities with no prior net use. In Côte d'Ivoire, it started roughly two years after the most recent mass net campaign, when net coverage had already declined. The trials measured mosquito densities before and after intervention, then tracked whether reductions in adult mosquitoes translated to fewer malaria cases.

The entomological results were striking. In Côte d'Ivoire, larval densities fell by 98 percent, and adult mosquito densities by 81 percent. In Kenya, the reductions were similarly substantial. Yet when researchers fitted mathematical models of malaria transmission to these mosquito counts, they discovered something unexpected: the same absolute reduction in mosquito numbers produced different epidemiological outcomes across villages. A village that lost 500 adult mosquitoes might see malaria prevalence drop by 30 percent, while another village losing the same number might see only a 10 percent decline. The variability was not random noise—it reflected real differences in how transmission works from place to place.

Multiple factors drove this heterogeneity. In Kenya, some villages had higher net use than others, and nets provide protection that compounds the effect of fewer mosquitoes. Microclimate differences altered how fast mosquito larvae developed from egg to adult—if conditions favored rapid development in one village, larvae could pupate and emerge faster than the weekly or fortnightly larvicide applications could suppress them. Terrain mattered too. In some valleys, permanent swamps harbored mosquito populations that were harder to reach; in others, habitats were more accessible and easier to treat comprehensively. The surrounding landscape also influenced whether adult mosquitoes migrated into treated villages from untreated areas nearby. In Côte d'Ivoire, a different puzzle emerged: baseline malaria incidence was higher in some villages despite similar or even lower mosquito densities. This paradox—high burden with low vectors, or low burden with high vectors—has been documented elsewhere, particularly around rice irrigation schemes where better housing and higher incomes near productive agricultural areas correlate with lower malaria despite abundant mosquitoes.

The researchers used mathematical models to explore why the same intervention produced variable results. The models showed that larviciding's benefit depends partly on how much transmission occurs outdoors, beyond the reach of bed nets. In high-transmission settings with substantial outdoor biting, reducing mosquitoes through larviciding offers greater relative protection. The models also revealed a counterintuitive finding: large absolute reductions in mosquito numbers can produce small or undetectable relative reductions in transmission if the baseline mosquito population is enormous. In the Gambian floodplains, for instance, an earlier attempt at larviciding failed because breeding sites were so numerous and terrain so difficult that teams could not treat enough of them to meaningfully suppress the population.

These findings complicate the path to scaling up larval source management across Africa. The trials demonstrate that the approach works—all villages receiving larviciding benefited—but effectiveness is not uniform. Predicting impact requires understanding local ecology, baseline malaria burden, net use patterns, and vector behavior. Large cluster-randomized trials, the gold standard for testing interventions, may be impractical given the high variability in effect sizes across sites. Instead, researchers suggest that well-powered observational studies combined with mathematical modeling could generate the evidence needed to guide where and how to deploy LSM. New technologies that can treat larval habitats without first identifying every breeding site may help reduce the effort required. The path forward depends on good planning, local expertise in eco-epidemiological conditions, and sustained financing—but the evidence suggests that where conditions align, larviciding can deliver meaningful reductions in malaria burden.