Applying Wolbachia to Eliminate Dengue

Primary objective:

To assess the efficacy of community-based deployments of Wolbachia-infected Aedes aegypti mosquitoes in reducing the incidence of symptomatic, virologically-confirmed dengue cases of any severity in Yogyakarta residents aged 3-45 years in release areas, relative to non-release areas.

Secondary objectives:

- To measure the efficacy of the Wolbachia method against each of the four DENV serotypes.

- To measure the efficacy of the Wolbachia method in reducing the incidence of symptomatic virologically confirmed Zika virus and chikungunya virus infection in release areas, relative to non-release areas

- To quantify the level of human mobility within Yogyakarta City, and estimate the proportion of residents' exposure time that they spend outside the treatment arm to which they were randomised

- To determine whether community-based deployment of Wolbachia-infected Ae. aegypti mosquitoes reduces the abundance of wild-type Ae. aegypti adults, or alternatively, alters the abundance of adults from Aedes species other than Ae. aegypti (e.g. Ae. albopictus)

Study setting:

The study will be conducted in Yogyakarta City and Bantul District, both located in the province of Yogyakarta Special Region, Indonesia.The study site is 26 km2 in size, including 24 km2 within Yogyakarta City, and 2km2 in the adjacent Bantul District. The total population of the study area is approximately 350,000.

Study design:

A cluster randomised trial with a test-negative design will be conducted. The study site will be divided into 24 clusters. The intervention will be allocated using constrained block randomisation with a parallel 1:1 assignment of intervention and control.

The intervention is the deployment of Wolbachia-infected Aedes aegypti mosquitoes. Wolbachia deployments will be conducted in intervention clusters with the aim of achieving Wolbachia establishment (>80% mean Wolbachia prevalence in trapped mosquitoes) throughout intervention areas within one year.

The impact of Wolbachia deployments on dengue incidence will be assessed by comparing the exposure distribution (probability of living in a Wolbachia-treated area) among virologically-confirmed dengue cases presenting to a network of public primary clinics (Puskesmas), against the exposure distribution among patients with febrile illness of non-arboviral aetiology presenting to the same network of clinics in the same temporal windows. Dengue cases and arbovirus-negative controls will be sampled concurrently from within the population of patients presenting with febrile illness to the study clinic network, with case or control status classified retrospectively based on the results of laboratory diagnostic testing.

A re-estimation of sample size requirements was conducted in January 2019 after one year of recruitment. The initial power calculation used 1000 dengue cases and 4000 non-dengue controls allocated to each cluster based on historical proportions of dengue cases and other febrile illnesses, assuming no variation in the proportion of cases by cluster. This method was found to overestimate power for small samples by not taking into account randomness in the sampling. The sample size re-estimation included power estimates for 200, 400, 600, 800 and 1000 dengue cases with 4 times as many controls allocated to each cluster by sampling from a multinomial distribution, which incorporated added randomness by allowing the proportion of cases allocated to each cluster to vary across simulations. The re-estimation found that 400 dengue cases plus four times as many controls would be sufficient to detect a 50% reduction in dengue incidence with 80% power.

Participant selection:

Participants will be enrolled from within the population of patients presenting with undifferentiated fever of 1-4 days duration, to one of the participating local health clinics (Puskesmas).

All patients meeting the inclusion criteria will be invited to participate in the study. From baseline historical data we expect approximately 5000 participants per annum to be enrolled.

Enrolment will continue for up to 36 months.

Analysis plan:

Permutation tests and standard regression models will be used to estimate the relative risk of dengue in Wolbachia-treated versus untreated clusters, accounting for the non-independence of study participants resident in the same intervention cluster, and temporal matching of dengue cases and test-negative controls.

The intention-to-treat analysis will consider Wolbachia exposure as binary depending on the allocation of the cluster of residence.

The per-protocol analysis will consider Wolbachia exposure as a continuous weighted index based on Wolbachia prevalence in trapped mosquitoes in the cluster of residence, either with or without weighting for time spent in other clusters visited during the ten days prior to illness onset.