Mosquito-disseminated pyriproxyfen for mosquito-borne disease control in Belo Horizonte, Brazil: a pragmatic, before-after control-intervention paired-series trial.

Background: Mosquitoes transmit important human pathogens, including dengue virus, but are notoriously hard to control. Mosquito-disseminated pyriproxyfen (MDPPF) uses the mosquitoes themselves to transfer particles of pyriproxyfen, a potent larvicide and pupicide, from lure dissemination stations to untreated larval habitats. MDPPF can reduce mosquito densities, but possible epidemiological effects remain to be measured. We aimed to investigate whether MDPPF can help curb mosquito-borne disease transmission.

Methods: In this pragmatic, before-after control-intervention paired-series (BACIPS) trial conducted in Belo Horizonte, Brazil, municipal vector-control staff deployed, then serviced monthly (from November, 2017, to December, 2019), 2481 pyriproxyfen dissemination stations in a nine-neighbourhood cluster with a history of high dengue endemicity; nine adjacent neighbourhoods were designated as a buffer area, and the remaining 258 city neighbourhoods as the control area. The primary epidemiological outcome of the trial was dengue incidence. Based on official dengue-notification records broken down by week and neighbourhood (ie, week-neighbourhood case counts; N=265 162 cases in total) from Jan 1, 2016, to Dec 31, 2019, we estimated intervention effects on incidence using a BACIPS approach and negative-binomial generalised linear mixed models (GLMMs). Zika and chikungunya cases were too rare to be assessed with confidence.

Findings: Week-neighbourhood dengue incidence ranged from 0 to 379·5 cases per 10 000 residents, with epidemic outbreaks recorded in 2016 and 2019. Intention-to-treat, BACIPS-GLMM adjusted estimates indicate that MDPPF deployment was associated with a net 29% (95% CI 21-36; p=4·7 × 10^-10) average decrease of dengue incidence in intervention neighbourhoods and a net 21% (12-30; p=2·7 × 10^-5) average decrease in buffer neighbourhoods. In contrast, and due in part to larger uncertainties, average incidence rates were statistically indistinguishable across areas before the intervention (intervention area p=0·47; buffer area p=0·11) and across trial periods in control neighbourhoods (p=0·74). Hence, in the all-too-common scenario of a 100 000-case outbreak, public health managers could expect MDPPF to reduce the strain on the health-care system by at least about 29 000 (21 000-36 000) symptomatic cases.

Interpretation: Our results suggest that MDPPF can help prevent dengue under the many operational constraints of real-world vector-control interventions and despite incomplete coverage and potential dilution of intervention effects. MDPPF holds promise as an additional tool for dengue control.

Funding: Coordenação-Geral de Vigilância de Arboviroses, Secretaria de Vigilância em Saúde e Ambiente, Ministry of Health, Brazil, Secretaria Municipal de Saúde de Belo Horizonte, Brazil, and Fundação de Amparo à Pesquisa do Estado do Amazonas, Brazil.