Spatial exploration of the impact of cattle on risk of malarial infection using data from a cluster randomised trial

Introduction Malaria is a parasitic disease transmitted by Anopheles mosquitoes.1 Global efforts to control the disease led to a marked reduction in the disease burden from 2000 to 2015.1 However, progress has stalled since then, with the WHO reporting an increase of 52 million cases globally from 2015 to 2024.1 About 95% of malaria deaths occur in the African Region, where just over 75% of the deaths are among children under 5 years of age.1 The disruption caused by the COVID-19 pandemic to health systems worldwide can only partly explain the lost progress as there are many other factors hindering malaria control, including climate change, insecticide resistance, mosquito behavioural adaptations and antimalarial drug resistance, among others.1 Vector-related challenges are of special concern because previous successes in the fight against malaria have been largely attributed to the widespread implementation of vector control tools, mainly long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS).2 These core vector control tools work mostly against mosquitoes that bite humans indoors at night, which are the most efficient malaria vectors in most African settings.3 Nonetheless, most malaria vectors feed at least partially on animals (ie, they are partially zoophilic) and have important outdoor and crepuscular activity,4 being able to sustain residual malaria transmission even after good coverage and usage of all available vector control tools.3–5 Furthermore, the intensive use of LLINs and IRS has been associated with a shift towards outdoor, zoophilic behaviour.6–9 Considering this scenario, evidence suggests that the currently available tools will not be enough to control malaria and achieve elimination in all settings.10–12 Livestock keeping is an essential source of income and wealth, particularly in rural settings where malaria is more prevalent.13 Given the proclivities of some malaria vectors for animal hosts, there has been extensive exploration of the possibility of using livestock as zooprophylaxis for malaria, which means using the animals to attract mosquitoes, thus reducing the number of bites on humans. Some studies have found significant associations between livestock keeping and malaria reduction,14–16 supporting that approach. However, other studies have reported a significant increase in malaria prevalence related to owning cattle and/or pigs,17 18 and one study saw no association between malaria and livestock ownership at all.19 While mathematical models have explored the conditions under which a zooprophylaxis strategy might be efficacious,20 21 the two systematic reviews on the topic agree that there are context-specific key factors to consider, including the distance between where the humans sleep and where the animals are kept.22 23 The opposite effect of zooprophylaxis is zoopotentiation, where the presence of livestock enhances malaria transmission either by serving as an alternative and relatively unprotected blood source for mosquitoes, thus allowing to sustain the vector populations, or by contributing to the creation and sustainment of breeding sites where mosquitoes can reproduce, such as muddy pens full of puddles enriched with organic matter.24 The challenge with many studies investigating the role of livestock in malaria is that they examine the association based on livestock ownership at the household level, considering it similarly to other household-level variables such as housing materials or family assets. However, both zooprophylaxis and zoopotentiation require proximity of livestock to humans, which may not be directly associated with ownership. For example, in semiextensive livestock systems commonly found in malaria-endemic areas, families keep their livestock close to or within the homestead at night. As a result, where the houses are closely spaced, households that do not own livestock might still be impacted by the presence of animals owned by neighbours. Hence, it is perhaps more appropriate to shift the epidemiological framework and consider livestock as an environmental variable, similar to the presence of water bodies around the household, which impact vector life cycles and alter vector abundance. Livestock available as an alternative blood meal source for mosquitoes might contribute to changes in vector population size in a similar fashion. In other words, the impact of livestock on malaria transmission is more likely to depend on the number of animals living around humans than on the fact that the animals belong to any specific individual or household. In this study, we use data collected during a community-level randomised controlled trial evaluating a novel vector control tool for malaria in coastal Kenya25 to explore the role of livestock on the risk of malaria infection and what factors might contribute to a zooprophylactic or zoopotentiation effect. We hypothesise that cattle presence around households is more predictive of malaria transmission than cattle ownership. To test this, we divert from previous studies on the topic and implement a novel spatial approach that considers cattle density and cattle-to-human-ratio in increasing concentric areas around the household, regardless of who owns the animals. That is, livestock is explored as an environmental exposure rather than a family asset.