SIAM News Blog
SIAM News
Print

Controlling the Spread of Visceral Leishmaniasis with Deltamethrin-Impregnated Dog Collars

By Lina Sorg

Visceral leishmaniasis (VL), also known as kala-azar, is a vector-borne zoonotic disease characterized by irregular episodes of fever, weight loss, anemia, and enlargement of the spleen and liver. If left untreated, 95 percent of cases are fatal. VL is most prevalent in Brazil, parts of East Africa, and India, though many other areas of the world experience outbreaks as well (see Figure 1). In fact, the disease is endemic in 77 countries and constitutes a global problem. More than 600 million people worldwide are at risk for VL at any given time. There are between 50,000 and 90,000 diagnosed cases each year, up to 30,000 of which end in death. 

The sandfly, a tiny insect that lives in tropical or temperate regions around the world, is the primary vector of VL. And because dogs are the sandfly’s principal reservoir host, the presence of seropositive dogs increases the risk of VL infection among the human population. Numerous studies have explored the feasibility of culling dogs to control the spread of VL, though the effectiveness of this prospective method is controversial. While researchers agree that dogs inhibit VL mitigation efforts, the complete absence of dogs within a vulnerable community is unrealistic. During a minisymposium at the 2021 SIAM Conference on Applications of Dynamical Systems, which is taking place virtually this week, Mondal Hasan Zahid of the University of Michigan explored the efficacy of deltamethrin-impregnated dog collars (DIDCs) to manage VL. The deltamethrin insecticide kills the sandflies but does not harm dogs or humans, therefore eliminating the vector and serving as a control measure.

Figure 1. Status of endemicity of visceral leishmaniasis (VL) worldwide.

Zahid set out to model and compare two different scenarios: one in which dogs simply do not exist, and one in which dogs wear DIDCs. The presence of DIDCs has two opposite impacts. Although the insecticide causes a reduction in the sandfly population, the presence of dogs supplies the remaining sandflies with more blood-meal sources, which are necessary for their survival. Zahid sought to understand the net effect of these two contrasting outcomes.

He began by introducing earlier works that employed similar models to answer different questions about VL. For example, a 2016 model addressed the relationship between dogs, sandflies, and humans. In this model, infectious sandflies were able to infect both dogs and humans, and infectious dogs could in turn infect the sandlfy population. The authors also assumed that humans are dead-end hosts and cannot transmit the disease. A similar study in 2017 generated a comparable model that also assumed the inability of humans to cause infections.

Figure 2. General model of the three populations involved in the spread of visceral leishmaniasis (VL): dogs, sandflies, and humans. The solid lines indicate full compartments within the same population.

However, a clinical study in 2000 found that humans can in fact transmit pathogens to sandflies. Zahid therefore wanted to include this transmission in his general model of the three populations (see Figure 2). The left side of the model accounts for the dog population, the right side accounts for humans, and sandflies are on top. Both the dog and human populations cycle between “susceptible” \((S)\), “exposed” \((E\) or \(L)\), “infectious” \((I)\), and “recovered” \((R)\) before returning to \(S\) because immunity is not indefinite. Exposed humans and dogs can also move directly to \(R\), because certain studies have shown that a portion of these individuals develop immunity before becoming infectious.

Zahid’s model differs from existing models because infectious humans can transmit the disease. Another unique feature is that the encounter rate between the vector and host is a function of host availability. When defining the encounter rate, Zahid considered the required number of bites/blood meals for a sandfly, as well as the number of bites that a dog or a human might experience in a single day.

Figure 3. The basic reproduction numbers \((R_0)\) for three scenarios: dogs with deltamethrin-impregnated dog collars (DIDCs), unprotected dogs, and no dogs.
Next, Zahid removed the dog population from his model. He then contrasted this amended state (of only sandflies and humans) with the original case, comparing the basic reproduction numbers \((R_0)\) for three scenarios: dogs with DIDCs, unprotected dogs, and no dogs (see Figure 3). 39,445 individuals comprise the model’s population size, which is based on population data for villages in Brazil. The protected dog scenario yields the lowest number of cases and the lowest \(R_0\) value. And though the complete absence of dogs has the highest \(R_0\), it still produces fewer infections than an unprotected dog scenario because there are no contributions from the dog population. The unprotected dogs therefore cause more infections than the other two instances.

Ultimately, Zahid found that the use of DIDCs on dogs within an at-risk community yields fewer human infections of VL than culling the dog population or leaving the dogs unprotected. “Even if you cull all of the dogs, it is still not better than having protected dogs,” he said. Of course, these results depend on dog tolerance to sandfly bites and the efficacy of the DIDC. Although the percentage can vary based on different data sets and parameters, Zahid’s data indicates that at least 58 percent of dogs must have DIDCs to reduce the number of human cases of VL. The use of DIDCs could thus help curb disease spread in endemic areas.


This work has been published in Infectious Disease Modelling; the corresponding journal article is available online.


Lina Sorg is the managing editor of SIAM News.
blog comments powered by Disqus