Estimating the efficacy of community-wide use of systemic insecticides in dogs to control zoonotic visceral leishmaniasis: A modelling study in a Brazilian scenario

Abstract

Systemic insecticides in dogs have been suggested as a public health intervention to prevent human cases of Zoonotic Visceral Leishmaniasis (ZVL). But, currently there are no systemic insecticides for dogs registered against zoo-anthropophilic pool blood feeding phlebotomine flies. We predict the impact of community-wide use of systemic insecticide in dog populations as a public health measure to control transmission of Leishmania infantum to humans using a mathematical model. We developed a Susceptible-Exposed-Infected (SEI) compartmental model to describe L. infantum transmission dynamics in dogs, with a vectorial capacity term to represent transmission between L. infantum-hosting dogs via phlebotomine flies. For Infected (I) dogs two levels of infectiousness were modelled, high infectiousness and low infectiousness. Human incidence was estimated through its relationship to infection in the dog population. We evaluated outcomes from a wide range of scenarios comprising different combinations of initial insecticide efficacy, duration of insecticide efficacy over time, and proportion of the dog population treated (60%, 70% & 80%). The same reduction in human infection incidence can be achieved via different combinations of insecticide efficacy, duration and dog coverage. For example, a systemic insecticide with an initial efficacy of 80% and 6 months above 65% efficacy would require treating at least 70% of the dogs to reduce the human infection incidence by 50%. Sensitivity analysis showed that the model outcome was most sensitive to baseline values of phlebotomine fly daily survival rate and insecticide coverage. Community-wide use of systemic insecticides applied to the "L. infantum canine reservoir" can significantly reduce human incidence of L. infantum infection. The results of this mathematical model can help defining the insecticide target product profile and how the insecticide should be applied to maximise effectiveness.

Fig 1. Model representing the transmission dynamics of L. infantum.

(A) Compartmental model to calculate transmission between dogs: Susceptible (S)–Exposed (E)—Infectious (I). Proportion ρ of E dogs become highly infectious (I_HI), and 1-ρ become low infectious (I_LI). Vectorial capacity (C_D) represents the transmission of L. infantum among dogs. (B) Equations to estimate transmission from infected dogs to humans in the form of human incidence of ZVL (λ_H). Vectorial capacity (C_H) represents the transmission of L. infantum from dogs to humans. All the parameters included in C_D, C_H and λ_H are defined in Table 1.

Fig 2. Sand fly survival curve showing the continuous probability of sand fly survival.

Blue dashed-dotted line represents the sand fly survival after biting a dog treated with a systemic insecticide of 80% efficacy where only 20% of the sand flies survive after 2 days (black triangle). Black line is the baseline sand fly mortality reported by Dye, 1996. Red dotted line represents the lower bound used in the sensitivity analysis. Green dashed line represents the upper bound used in the sensitivity analysis. Black triangles represent survival 2 days after biting on dogs. Black squares represent survival of L. infantum extrinsic incubation period (7 days).

Fig 3. Reduction of human incidence of L. infantum infection.

Scenario of mass application of systemic insecticides to dogs. Dog coverage: 80% (A), 70% (B) and 60% (C). Insecticide efficacy (horizontal axis) is represented by the increase in sand fly mortality caused by the insecticide (μ_T(0)). Decay in the insecticide efficacy occurs at a constant rate per day (vertical axis). Contour curves mark 5% changes in human incidence.

Fig 4. Tornado plot showing the sensitivity of different parameters on the reduction in human incidence of L. infantum infection in the model.