Dengue Suppression by Male Wolbachia-Infected Mosquitoes
Published February 11, 2026
N Engl J Med 2026;394:1175-1183
DOI: 10.1056/NEJMoa2503304
Abstract
Background
Wild-type female Aedes aegypti mosquitoes that mate with male A. aegypti mosquitoes that have been infected with the wAlbB strain of Wolbachia pipientis bacteria produce nonviable offspring owing to cytoplasmic incompatibility. Repeated releases of wolbachia-infected males can potentially suppress wild-type mosquito populations and reduce the risk of dengue virus infection.
Methods
We conducted a trial involving the release of male A. aegypti mosquitoes infected with the wAlbB strain of wolbachia bacteria for the control of dengue in Singapore, a tropical city-state. In this cluster-randomized trial with test-negative controls, we divided 15 geographic population clusters into two groups: 8 clusters received deployments of male wolbachia-infected mosquitoes (intervention clusters) and 7 clusters received no deployments (control clusters). The primary end point was the diagnosis of symptomatic dengue virus infection of any severity caused by any serotype of the virus, as measured by the odds ratio for the distribution of wolbachia exposure among laboratory-confirmed reported dengue cases as compared with test-negative controls.
Results
A total of 393,236 residents lived in the intervention clusters, and 331,192 lived in the control clusters. Adult wild-type A. aegypti populations were suppressed across the intervention clusters. The baseline average abundance of the mosquitoes (number of adult female mosquitoes trapped divided by number of traps) was 0.18 and 0.19 in the intervention and control clusters, respectively; from 3 months after the initiation of the intervention until the end of the 24-month trial period, the average abundance was 0.041 and 0.277, respectively. In the intention-to-treat analysis at 6 months or more, the percentage of residents in the intervention clusters who were dengue-positive was lower than that in the control clusters (354 of 5722 tests [6%] vs. 1519 of 7080 tests [21%]). The protective efficacy of the intervention, calculated as (1−odds ratio)×100, ranged from 71 to 72% with 3 to 12 months or more of wolbachia mosquito exposure, as represented by odds ratios of 0.28 to 0.29.
Conclusions
Release of sterile wolbachia-infected male A. aegypti mosquitoes reduced vector populations and the risk of dengue infection in Singapore. (Funded by the Singapore Ministry of Finance and others; ClinicalTrials.gov number, NCT05505682.)
Dengue is a growing vectorborne disease globally. Because of increasing urbanization and climate change, the number and magnitude of dengue virus outbreaks in the past decade have grown steadily, which has led to increased morbidity and mortality worldwide.
However, the arsenal for dengue mitigation remains limited. For example, the CYD-TDV vaccine will cease production in coming years,1 and other vaccines such as TAK-003 and Butantan-DV have not yet shown high efficacy against all dengue serotypes.2,3 The efficacy of vector-control tools — such as community-based source reduction4 and use of spatial repellents5 — has been tested in a limited number of randomized, controlled trials. However, these trials did not show sustained epidemiologic efficacy in reducing dengue cases or the risk of dengue infection. In addition, these trials were conducted primarily in pediatric populations, which limits their applicability to at-risk settings that consist predominantly of adults. The use of spatial repellents showed modest protective efficacy against dengue, which ranged from 20 to 40% in treated locations.5 However, the use of chemicals may have off-target effects on other species and can disrupt the natural environment.
In recent years, Aedes aegypti mosquitoes that have been infected with Wolbachia pipientis bacteria have been touted as a viable strategy for reducing the transmission of dengue virus. One approach is to release both female and male A. aegypti wolbachia-infected mosquitoes to replace the wild-type mosquito populations with a wolbachia-infected population that has reduced potential for dengue transmission. In a cluster-randomized trial, investigators found that this intervention reduced the risk of dengue in areas of Yogyakarta, a city on the Indonesian island of Java.6 However, the protective efficacy that this intervention can confer has varied across different settings.7,8
Another approach is the use of wolbachia-mediated incompatible insect technique–sterile insect technique (IIT-SIT). This method involves the release of only male wolbachia-infected A. aegypti mosquitoes that have been irradiated to reduce the risk of the establishment of wolbachia-infected mosquitoes in the release sites because of the unintentional release of fertile females, which can happen owing to imperfect sex-sorting.9 Because of cytoplasmic incompatibility,10,11 offspring of wolbachia-infected males and uninfected wild-type females are not viable. Preliminary evidence suggests that repeated releases of wolbachia-infected males can drastically suppress wild-type mosquito populations12 and thus reduce dengue virus transmission.13,14 However, data are needed regarding the epidemiologic efficacy of IIT-SIT in randomized, controlled trials.
Here, we report the results of a cluster-randomized, controlled trial that assessed the efficacy of IIT-SIT releases of irradiated male wolbachia-infected mosquitoes in reducing the risk of dengue. The trial builds on baseline entomologic and epidemiologic field studies in the same geographic setting.12-14
Methods
Trial Design and Oversight
This unblinded, test-negative trial was conducted in Singapore, a tropical city-state, led by the Environmental Health Institute of the National Environment Agency. The trial relied on a nationally representative, passively monitored database of patients (ranging in age from newborn to 104 years old) who had been tested for suspected dengue virus infection. For many years, doctors across Singapore have been guided by the Ministry of Health to test for dengue virus in patients with compatible symptoms. No participants were enrolled in the trial. The protocol was published previously15,16 and is available with the full text of this article at NEJM.org.
The trial was approved by the Singapore Ministry of Health and the Dengue Expert Advisory Panel of the National Environment Agency. The trial data were analyzed by independent trial statisticians. The funders had no role in the analysis of the data, in the preparation or approval of the manuscript, or in the decision to submit the manuscript for publication. The authors vouch for the completeness and accuracy of the data and for the fidelity of the trial to the protocol.
This project was exempted from formal bioethics review, as advised by the Ministry of Health. All laboratory tests were performed for clinically directed reasons, and the data from these tests were routinely collected as part of the national dengue surveillance program under the Infectious Diseases Act legislation, which exempts the trial from the need for informed consent.
Randomization
The baseline characteristics of the trial geographic population clusters are shown in Table 1. In brief, the trial sites consisted of urban areas totaling 12.0 km2 with an overall population of an estimated 724,428 residents of high-rise public housing apartments. The trial site was subdivided into 15 clusters, with each cluster averaging approximately 0.8 km2 in size. Each cluster had geographic borders that would minimize the dispersal of mosquitoes between clusters. Man-made or natural borders such as major roads, highways, and bodies of water were used to delineate cluster boundaries to limit spillover of wolbachia-infected mosquitoes from the intervention clusters to the control clusters, as well as the migration of wild-type mosquitoes into clusters where possible. In the absence of such borders, adjacent areas within a 300-m radius were designated as buffer release areas in case the cluster was designated for intervention.
Table 1
| Variable | Unweighted Data | Weighted Data† | |||
|---|---|---|---|---|---|
| Intervention Clusters | Control Clusters | Intervention Clusters | Control Clusters | SMD‡ | |
| Historical data | |||||
| Dengue incidence per 100,000 population per year: 2015–2021 | 183.9 | 188.4 | NA | NA | NA |
| Positive dengue test: 2016–2021 — % | 18.0 | 19.9 | NA | NA | NA |
| Resident demographic data | |||||
| Age on date of testing — yr | 44.9±24.4 | 46.0±24.6 | 45.8±24.4 | 45.8±24.61 | <0.01 |
| Male sex — % | 51.5 | 50.7 | 50.9 | 50.9 | <0.01 |
| Cluster data§ | |||||
| Vegetation index | 0.32±0.04 | 0.32±0.04 | 0.33±0.04 | 0.33±0.04 | <0.01 |
| Location <300 m from body of water — % | 0.2±0.3 | 0.3±0.4 | 0.3±0.4 | 0.3±0.4 | <0.01 |
| Height of residence — m | 35.3±6.6 | 41.0±13.0 | 36.8±6.8 | 36.8±8.9 | <0.01 |
| Price of residence — S$¶ | 482,933±86,754 | 538,554±129,000 | 500,272±88,553 | 500,272±98,600 | <0.01 |
| Length of drainage network within spatial unit — m | 57.5±149.5 | 68.7±157.2 | 72.2±181.3 | 72.2±151.1 | <0.01 |
| Total vegetation — % | 0.04±0.09 | 0.02±0.03 | 0.02±0.04 | 0.02±0.03 | <0.01 |
| Climate conditions‖ | |||||
| Mean temperature, lag 3 — °C | 28.0±1.1 | 28.1±1.0 | 28.0±1.1 | 28.0±1.1 | <0.01 |
| Mean wind speed — km/hr | 8.0±2.3 | 7.9±2.3 | 7.9±2.3 | 7.9±2.4 | <0.01 |
| Relative humidity — % | |||||
| Current | 81.8±3.0 | 81.3±2.9 | 81.5±3.0 | 81.5±3.0 | <0.01 |
| Lag 4 | 81.8±3.0 | 81.4±3.0 | 81.6±3.0 | 81.6±3.1 | <0.01 |
| Weekly mean daily rainfall — mm | |||||
| Current | 7.6±7.2 | 7.6±6.5 | 7.6±7.0 | 7.5±6.5 | <0.01 |
| Lag 1 | 7.6±7.2 | 7.8±6.6 | 7.7±7.2 | 7.7±6.7 | <0.01 |
| Lag 2 | 7.7±7.4 | 8.0±6.8 | 7.8±7.5 | 7.8±6.7 | <0.01 |
| Lag 3 | 7.6±7.1 | 8.0±6.7 | 7.78±7.2 | 7.8±6.6 | <0.01 |
| Lag 4 | 7.5±6.9 | 7.8±6.7 | 7.7±7.1 | 7.7±6.4 | <0.01 |
*
Plus–minus values are means ±SD. Data for the intervention group include all the residents who were tested for dengue and had any duration of exposure to wolbachia-infected mosquitoes. Data for the control group include all the residents for whom information was available regarding the date of testing, starting on the first day of release of wolbachia-infected mosquitoes in the trial. NA denotes not applicable.
†
Overlap weights were used to weight baseline characteristics between the two trial clusters. Overlap weights were computed as 1minus the propensity score for observations in the intervention clusters and equal to the propensity score for observations in the control clusters. Propensity scores were obtained by training a logistic-regression analysis with randomization to the intervention clusters or control clusters as the dependent variable of interest and anthropogenic or environmental characteristics as the explanatory variables.
‡
A standardized mean difference (SMD) of less than 0.05 indicates good balance in characteristics between the comparator groups.
§
Clusters were reported to have negligible forest or grass cover.
¶
Values are reported in Singapore dollars (S$), with S$1.00 equal to $0.77 in U.S. dollars.
‖
Lag values for temperature, humidity, and rainfall are provided because of the delayed effect of meteorologic variables on mosquito survival or fecundity. These values were calculated as the recorded measurements of meteorologic data 1 to 4 weeks before the dengue test date.
Clusters (including the buffer areas) were kept at least 700 m apart. Of the 15 clusters, 8 were randomly assigned to receive open-label wolbachia-infected mosquitoes (intervention clusters), and 7 clusters were assigned to receive no mosquito release (control clusters). No placebo intervention was used in the control clusters. Constrained randomization was conducted to prevent chance imbalance in the historical risk of dengue at baseline in the intervention and control clusters. Details regarding the randomization methods are provided in the Supplementary Appendix, available at NEJM.org.
Deployment and Monitoring of Mosquitoes
The half-life of released male A. aegypti mosquitoes is 4 days, so the protocol called for twice-weekly release of one to six mosquitoes per human resident per week on weekday mornings between 6:30 a.m. and 11 a.m. in the intervention clusters. Male wolbachia-infected mosquitoes were also released twice a week on the ground floors of high- and low-rise housing estates within the designated buffer areas. If releases in buffer areas could not be performed (e.g., in proximity to schools), releases were conducted at the boundary between such areas and the intervention sites.
We generated a localized wolbachia-infected A. aegypti line by crossing a Singapore wild-type line with a wolbachia-infected line imported from the United States over six generations, which resulted in complete maternal transmission and complete cytoplasmic incompatibility. The production of male wolbachia-infected mosquitoes consisted of the separation of male A. aegypti mosquitoes, which were then exposed to low-dose irradiation to sterilize any residual females. This process prevents inadvertent release of fertile females and hence also the establishment of wolbachia-infected mosquitoes in the release sites, while maintaining male mating competitiveness. (Details regarding this process are provided in the Supplementary Appendix.)
Intervention clusters first received the wolbachia-infected mosquitoes in 2022 during epidemiologic weeks 30 to 37. The trial was concluded in epidemiologic week 37 in 2024. Adult A. aegypti populations in intervention and control clusters were monitored with the use of Gravitraps (containers that are designed to attract and capture female mosquitoes), with an average of six Gravitraps deployed per apartment block. The abundance of mosquitoes17 was quantified by means of the weekly Gravitrap A. aegypti index (GAI), which was defined as the total number of female adult A. aegypti mosquitoes that were caught in functional Gravitraps divided by the number of functional Gravitraps.
Dengue Surveillance
Epidemiologic efficacy was assessed through the use of two nationally representative datasets. In the first set, data regarding residents with suspected dengue illness who were tested for dengue were collated from all major diagnostic laboratories through primary care clinics and hospitals. The second dataset was collated from the national dengue surveillance system, in which all cases of dengue are legally mandated to be reported to the Ministry of Health. When doctors from primary care clinics or hospitals across Singapore identify patients with suspected dengue, blood samples are sent for dengue virus tests through a national network of diagnostic laboratories. Under the Infectious Diseases Act, the reporting of all laboratory-confirmed cases of dengue virus infection to the national surveillance program is legally mandated.
Resident Eligibility and Inclusion
During the trial period from epidemiologic week 30 of 2022 to epidemiologic week 37 of 2024, samples that had been obtained from 189,670 residents with suspected dengue virus infection (dengue-testing database) were analyzed at the Environmental Health Institute, at hospital laboratories and major commercial diagnostic laboratories, through consultations at primary care clinics or hospitals. More than 30% of dengue-positive samples were serotyped at the trial sites during the trial period. At that time, the predominant serotype was dengue virus type 3. (Full details are provided in the Supplementary Appendix.)
Information that we collected regarding test-positive dengue cases represented 85.9% (29,370 of 34,208) of all the dengue cases that were reported to the national dengue notification database during the trial period. We excluded results from residents who had more than one residential address in different control or intervention clusters, those who had been tested at multiple laboratories with conflicting dengue results, and those who had no corresponding residential address. We also excluded residents who had an address at an intervention site at the time of the test but had not been exposed to the intervention for prespecified monthly periods, according to exposure criteria described below. We also excluded residents who had incomplete demographic data. For residents who had undergone repeated testing, we classified residents as having a positive test if they had one positive test out of multiple tests within 4 weeks of each other; otherwise, we used the latest negative result (Figure 1). Investigators had access to residents’ testing status and cluster assignment, with masking of personal details.
Figure 1
Procedures, Diagnostic Investigations, and Classifications
All the residents with suspected dengue virus infection were tested with a reverse-transcriptase–quantitative polymerase-chain-reaction assay, a dengue nonstructural protein 1 (NS1) assay, or an IgM antibody or IgM–NS1 duo assay to detect dengue virus in blood samples.15,18 Residents were considered to have a positive dengue test if the sample was positive on any of these assays. If residents underwent more than one test, they were considered to be positive for dengue if any of the tests had positive results. Patients were classified as negative controls if the tested sample was negative. Data regarding place of residence and demographic details were collected at the time of testing.
Primary End Points
The study had two primary end points: dengue risk and dengue incidence. Only the first of these end points — involving the diagnosis of symptomatic dengue of any severity caused by any serotype — is described here. This end point was measured by the odds ratio of exposure to wolbachia-infected mosquitoes among residents with laboratory-confirmed dengue as compared with dengue-negative controls. Comparisons were made between intervention clusters and control clusters.
Statistical Analysis
We determined that 400 dengue-positive residents and 1600 dengue-negative controls would be needed to provide the trial with 80% power to detect a risk of dengue that was 50% lower in the intervention clusters than in the control clusters. A full description of the sample-size calculations was published previously.15,16 The statistical analysis plan is described in full detail in the trial protocol.
For the first primary end point in the intention-to-treat analysis, we considered exposure to wolbachia-infected mosquitoes as a binary classification according to whether residents had been exposed for 0, 3, 6, 9, 12, or more months at the time that they had reported for testing. Residence was defined as the reported place of residence at the time of testing. The effect of the intervention was estimated with weighted logistic regression with overlap weights to account for post hoc imbalance in spatiotemporal characteristics. The null hypothesis was that the odds of residing in an intervention cluster would be the same among residents who tested positive for dengue and those who tested negative. Protective efficacy of the intervention was calculated as (1 minus odds ratio)×100.
Results
Suppression of Mosquito Populations
Suppression of adult wild-type A. aegypti populations was shown across all intervention clusters, with reductions in the GAI measure over time. Preintervention GAI levels averaged 0.18 and 0.19 in the intervention and control clusters, respectively. From 3 months after the initiation of the intervention to the end of the 24-month trial period, the average GAI decreased to 0.041 in the intervention clusters, as compared with an increase to 0.277 in the control clusters (Figure 2).
Figure 2
Clusters and Residents
During the trial period, after inclusion and exclusion criteria had been met, testing for suspected dengue virus infection was performed in 8245 residents in the intervention clusters and in 10,344 residents in the control clusters (Figure 1). The mean (±SD) age of the residents was 45.5±24.5, and 51.5% of the residents were male (Table 1). The age, sex, and other sociodemographic characteristics of the residents were well matched in the intervention and control clusters.
Primary End Point
In the intention-to-treat analysis at 6 months or more, the percentage of residents in the intervention clusters who were dengue-positive was lower than that in the control clusters (354 of 5722 tests [6%] vs. 1519 of 7080 tests [21%]). The percentage of residents who tested positive for dengue viral infection was lower in the intervention clusters than in the control clusters across all periods of exposure to the wolbachia-infected mosquitoes. During the trial period, the percentages of positive cases ranged from 6 to 10% in the intervention clusters, as compared with 21 to 23% in the control clusters. The estimated overall protective efficacy with 3 to 12 months or more of the intervention ranged from 71 to 72% (represented by odds ratios of 0.28 to 0.29).
The protective efficacy remained fairly constant across all exposure durations (Table 2). The protective efficacy was consistent across all age and sex subgroups, as well as across each calendar year (Table 3). Similarly, the protective efficacy in general remained stable as the exposure time increased in all subgroups and remained consistent after test results determined with IgM assays were removed from the analysis.
Table 2
| Exposure Period† | Protective Efficacy (95% CI) | Positive Test for Dengue | |
|---|---|---|---|
| Intervention Clusters | Control Clusters | ||
| percent | percent (no./total no.) | ||
| 0+ mo | 65 (47–77) | 10 (787/8245) | 23 (2402/10344) |
| 3+ mo | 71 (58–81) | 7 (457/6743) | 21 (1717/8125) |
| 6+ mo | 72 (58–82) | 6 (354/5722) | 21 (1519/7080) |
| 9+ mo | 71 (55–82) | 7 (317/4668) | 23 (1403/6118) |
| 12+ mo | 71 (51–84) | 7 (256/3724) | 21 (1037/4986) |
*
Protective efficacy was computed as (1minus odds ratio)×100, in which the odds ratio was estimated by means of logistic regression with overlapping weights.
†
The exposure period refers to the minimum amount of time a resident was exposed to wolbachia interventions at the point of testing.
Table 3
| Subgroup | Protective Efficacy | ||||
|---|---|---|---|---|---|
| Exposure for 0+ Months | Exposure for 3+ Months | Exposure for 6+ Months | Exposure for 9+ Months | Exposure for 12+ Months | |
| percent (95% CI) | |||||
| Age | |||||
| 0 to 20 yr | 63 (42–77) | 67 (50–79) | 67 (44–80) | 66 (36–81) | 59 (32–77) |
| 20 to 65 yr | 64 (45–77) | 70 (56–81) | 71 (56–82) | 72 (57–82) | 72 (54–83) |
| >65 yr | 72 (59–83) | 76 (64–86) | 77 (64–86) | 75 (59–87) | 77 (53–90) |
| Sex | |||||
| Female | 68 (48–81) | 74 (61–84) | 76 (61–85) | 77 (63–86) | 76 (58–88) |
| Male | 63 (45–75) | 68 (55–79) | 68 (54–78) | 66 (49–79) | 67 (45–81) |
| Year | |||||
| 2022 | 52 (11–77) | 68 (41–84) | NA | NA | NA |
| 2023 | 62 (33–76) | 62 (33–76) | 62 (38–76) | 60 (24–76) | 50 (21–66) |
| 2024 | 77 (55–88) | 76 (54–87) | 76 (53–87) | 76 (51–87) | 76 (51–87) |
*
NA denotes not applicable.
Discussion
During the trial period, the abundance of A. aegypti mosquitoes was greatly reduced in the intervention clusters as compared with the control clusters. The reduced number of mosquitoes resulted in a reduction of 71 to 72% in the risk of dengue virus infection among residents of the intervention clusters with at least 3 months of exposure. Protective efficacy was shown across all calendar years and in all age and sex subgroups, which indicated consistent biologic replication of the effect of the intervention.
This trial leveraged both extensive vector and dengue surveillance systems. A nationwide mosquito surveillance system with a consistent monitoring protocol was used to assess entomologic efficacy. Epidemiologic efficacy was assessed through the use of two nationally representative databases: the dengue-testing database and the national dengue surveillance system. Confirmatory diagnostic testing of suspected dengue cases with highly sensitive and specific tests is widely available across health care settings in Singapore. These data enabled a minimization of selection bias and detection of dengue among residents who sought health care and had suspected dengue infection. The randomization design that focused on geographic clusters rather than individual residents allowed for similar levels of historical risk of dengue infection in the trial population. Retrospective balancing or adjustment for many environmental and anthropogenic confounders that may potentially drive the risk of dengue was performed, and no evidence was found that such factors may have influenced the estimated efficacy of the intervention.
Although the trial was conducted under a backdrop of low population-level seroprevalence of antibodies against dengue virus and low acquisition of dengue infection,19 in locations where IIT-SIT was used, a reduction in the risk of dengue was reported. The efficacy results in our trial are also consistent with the body of laboratory, modeling, and field literature. IIT-SIT has been found to drive the suppression of wild-type A. aegypti mosquitoes in long-term field trials in Singapore, as well as in another pilot trial conducted in Mexico,9,12,20 although entomologic effectiveness was found to be lower in a stand-alone IIT trial conducted in Puerto Rico.21 In a stand-alone SIT program in Brazil, an estimated 89.4% reduction in live progeny of field A. aegypti mosquitoes and 97% reduction in dengue incidence were reported.22 Reductions in wild-type A. aegypti populations caused by releases of wolbachia-infected mosquitoes may consequentially drive the reductions in both the risk14,23 and the incidence13 of dengue virus infection.
This mosquito-based suppression technology has several potential advantages. First, although our trial showed protective efficacy against dengue (the only aedes-borne disease in constant endemic circulation in Singapore), a similar protective efficacy would be expected against other aedes-borne diseases transmitted by this mosquito — including Zika, chikungunya, and yellow fever — through the suppression of vector populations. Second, although dengue virus type 3 was the predominant serotype in circulation during the trial period, consistent efficacy would be expected against other dengue serotypes because the intervention targets the adult A. aegypti populations. Third, in our trial, consistent entomologic efficacy and protective efficacy against dengue was shown and maintained in a densely populated region with a large at-risk population (>700,000 residents). Fourth, baseline studies at pilot sites showed public acceptance of the intervention, with 77% of residents indicating support for releases, 18% reporting neutral opinion, and only 4% expressing lack of support.24 Fifth, our findings mirror those of earlier studies12,13,25 in which 100% cytoplasmic incompatibility was induced by the same wolbachia strain that was used in our trial. The implementation of a wolbachia suppression program at a national level has been shown to be cost-effective at an efficacy threshold of 40% for the intervention in a similar trial setting,26 despite a higher up-front implementation cost than other approaches owing to the need for sustained releases, sex separation, and irradiation of wolbachia mosquitoes before release. Previous modeling studies have shown that program-resource requirements can be significantly reduced by scaling down and redistributing capacity without compromising the suppressive efficacy of IIT-SIT.27
Our trial has several limitations. In performing an intention-to-treat analysis, we considered that exposure to treated mosquitoes would be based on place of residence, even though trial residents could be exposed to dengue in locations outside the designated clusters. The trial was conducted against a backdrop of extensive and integrated vector control in the trial setting, so efficacy results may not be generalizable to other locations. Although care was taken to reduce the migration of wild-type mosquitoes into the trial areas through geographic borders or buffer sites, such migration cannot be ruled out and may attenuate the estimated protective effects.
Release of sterile wolbachia-infected male A. aegypti mosquitoes may be a method for the control of dengue by reducing both vector populations and the risk of dengue virus infection. The technology can complement conventional approaches and vaccination to further reduce and potentially eliminate dengue transmission, along with possibly other aedes-borne diseases.
Notes
This article was published on February 11, 2026, at NEJM.org.
A data sharing statement provided by the authors is available with the full text of this article at NEJM.org.
Supported by the Singapore Ministry of Finance, the Ministry of Sustainability and the Environment, the National Environment Agency, and the National Robotics Program; and by a start-up grant (to Jue Tao Lim) from the Ministry of Education.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.
We thank the members of the Dengue Expert Advisory Panel of the National Environment Agency (Tim Barkham, Christl Donnelly, Neil Ferguson, Duane Gubler, Ary Hoffmann, and Chia Kee Seng) for their advice on the trial design and analysis; Zhiyong Xi (of Michigan State University) for providing the mosquitoes infected with the AlbB strain of wolbachia bacteria, which was used to generate the AlbB-SG line essential to this trial; and the members of the Production, Release, Partnerships, and Program Management teams of the Environmental Health Institute for their technical and program support.
Supplementary Material
References
1.
Arredondo-García JL, Hadinegoro SR, Reynales H, et al. Four-year safety follow-up of the tetravalent dengue vaccine efficacy randomized controlled trials in Asia and Latin America. Clin Microbiol Infect 2018;24:755-763.
2.
Kallás EG, Cintra MAT, Moreira JA, et al. Live, attenuated, tetravalent butantan-dengue vaccine in children and adults. N Engl J Med 2024;390:397-408.
3.
Biswal S, Reynales H, Saez-Llorens X, et al. Efficacy of a tetravalent dengue vaccine in healthy children and adolescents. N Engl J Med 2019;381:2009-2019.
4.
Vanlerberghe V, Toledo ME, Rodríguez M, et al. Community involvement in dengue vector control: cluster randomised trial. BMJ 2009;338:b1959-b1959.
5.
Morrison AC, Reiner RC Jr, Elson WH, et al. Efficacy of a spatial repellent for control of Aedes-borne virus transmission: a cluster-randomized trial in Iquitos, Peru. Proc Natl Acad Sci U S A 2022;119(26):e2118283119-e2118283119.
6.
Utarini A, Indriani C, Ahmad RA, et al. Efficacy of Wolbachia-infected mosquito deployments for the control of dengue. N Engl J Med 2021;384:2177-2186.
7.
Ribeiro Dos Santos G, Durovni B, Saraceni V, et al. Estimating the effect of the wMel release programme on the incidence of dengue and chikungunya in Rio de Janeiro, Brazil: a spatiotemporal modelling study. Lancet Infect Dis 2022;22:1587-1595.
8.
Pinto SB, Riback TIS, Sylvestre G, et al. Effectiveness of Wolbachia-infected mosquito deployments in reducing the incidence of dengue and other Aedes-borne diseases in Niterói, Brazil: a quasi-experimental study. PLoS Negl Trop Dis 2021;15(7):e0009556-e0009556.
9.
Soh S, Ho SH, Ong J, et al. Strategies to mitigate establishment under the Wolbachia incompatible insect technique. Viruses 2022;14:1132-1132.
10.
Yen JH, Barr AR. New hypothesis of the cause of cytoplasmic incompatibility in Culex pipiens L. Nature 1971;232:657-658.
11.
Sinkins SP. Wolbachia and cytoplasmic incompatibility in mosquitoes. Insect Biochem Mol Biol 2004;34:723-729.
12.
Bansal S, Lim JT, Chong C-S, et al. Effectiveness of Wolbachia-mediated sterility coupled with sterile insect technique to suppress adult Aedes aegypti populations in Singapore: a synthetic control study. Lancet Planet Health 2024;8(9):e617-e628.
13.
Lim JT, Bansal S, Chong CS, et al. Efficacy of Wolbachia-mediated sterility to reduce the incidence of dengue: a synthetic control study in Singapore. Lancet Microbe 2024;5(5):e422-e432.
14.
Lim JT, Mailepessov D, Chong C-S, et al. Assessing Wolbachia-mediated sterility for dengue control: emulation of a cluster-randomized target trial in Singapore. J Travel Med 2024;31(7):taae103-taae103.
15.
Ong J, Ho SH, Soh SXH, et al. Assessing the efficacy of male Wolbachia-infected mosquito deployments to reduce dengue incidence in Singapore: study protocol for a cluster-randomized controlled trial. Trials 2022;23:1023-1023.
16.
Lim JT, Mailepessov D, Chong C-S, et al. Update to: assessing the efficacy of male Wolbachia-infected mosquito deployments to reduce dengue incidence in Singapore. Trials 2024;25:400-400.
17.
Ong J, Chong C-S, Yap G, et al. Gravitrap deployment for adult Aedes aegypti surveillance and its impact on dengue cases. PLoS Negl Trop Dis 2020;14(8):e0008528-e0008528.
18.
Ho SH, Lim JT, Ong J, Hapuarachchi HC, Sim S, Ng LC. Singapore’s 5 decades of dengue prevention and control — implications for global dengue control. PLoS Negl Trop Dis 2023;17(6):e0011400-e0011400.
19.
Tan LK, Low SL, Sun H, et al. Force of infection and true infection rate of dengue in Singapore: implications for dengue control and management. Am J Epidemiol 2019;188:1529-1538.
20.
Martín-Park A, Che-Mendoza A, Contreras-Perera Y, et al. Pilot trial using mass field-releases of sterile males produced with the incompatible and sterile insect techniques as part of integrated Aedes aegypti control in Mexico. PLoS Negl Trop Dis 2022;16(4):e0010324-e0010324.
21.
Sánchez-González L, Crawford JE, Adams LE, et al. Incompatible Aedes aegypti male releases as an intervention to reduce mosquito population — a field trial in Puerto Rico. PLoS Negl Trop Dis 2025;19(1):e0012839-e0012839.
22.
de Castro Poncio L, Apolinário Dos Anjos F, de Oliveira DA, et al. Prevention of a dengue outbreak via the large-scale deployment of Sterile Insect Technology in a Brazilian city: a prospective study. Lancet Reg Health Am 2023;21:100498-100498.
23.
Lim JT, Mailepessov D, Chong CS, et al. Adjacent spillover efficacy of Wolbachia for control of dengue: emulation of a cluster randomised target trial. BMC Med 2025;23:184-184.
24.
Soh LT, Ong Z, Vasquez K, et al. A household-based survey to understand factors influencing awareness, attitudes and knowledge towards Wolbachia-Aedes technology. Int J Environ Res Public Health 2021;18:11997-11997.
25.
The Project Wolbachia–Singapore Consortium, Ching NL. Wolbachia-mediated sterility suppresses Aedes aegypti populations in the urban tropics. June 17, 2021 (https://www.medrxiv.org/content/10.1101/2021.06.16.21257922v1 preprint.
26.
Soh S, Ho SH, Seah A, et al. Economic impact of dengue in Singapore from 2010 to 2020 and the cost-effectiveness of Wolbachia interventions. PLOS Glob Public Health 2021;1(10):e0000024-e0000024.
27.
Lim PL, Cook AR, Bansal S, Chow JY, Lim JT. Wolbachia incompatible insect technique program optimization over large spatial scales using a process-based model of mosquito metapopulation dynamics. BMC Biol 2024;22:269-269.
Information & Authors
Information
Published In
Copyright
Copyright © 2026 Massachusetts Medical Society. All rights reserved.
For personal use only. Any commercial reuse of NEJM Group content requires permission.
History
Published online: February 11, 2026
Published in issue: March 26, 2026
Topics
Authors
Metrics & Citations
Metrics
Altmetrics
Citations
Export citation
Select the format you want to export the citation of this publication.