The epidemiology of imported and locally acquired dengue in Australia, 2012–2022

Abstract Background Dengue is the most important arboviral disease globally and poses ongoing challenges for control including in non-endemic countries with competent mosquito vectors at risk of local transmission through imported cases. We examined recent epidemiological trends in imported and locally acquired dengue in Australia, where the Wolbachia mosquito population replacement method was implemented throughout dengue-prone areas of northern Queensland between 2011 and 2019. Methods We analysed dengue cases reported to the Australian National Notifiable Disease Surveillance System between January 2012 and December 2022, and Australian traveller movement data. Results Between 2012 and 2022, 13 343 dengue cases were reported in Australia (median 1466 annual cases); 12 568 cases (94.2%) were imported, 584 (4.4%) were locally acquired and 191 (1.4%) had no origin recorded. Locally acquired cases decreased from a peak in 2013 (n = 236) to zero in 2021–22. Annual incidence of imported dengue ranged from 8.29/100 000 (n = 917 cases) to 22.10/100 000 (n = 2203) annual traveller movements between 2012 and 2019, decreased in 2020 (6.74/100 000 traveller movements; n = 191) and 2021 (3.32/100 000 traveller movements; n = 10) during COVID-19-related border closures, then rose to 34.79/100 000 traveller movements (n = 504) in 2022. Imported cases were primarily acquired in Southeast Asia (n = 9323; 74%), Southern and Central Asia (n = 1555; 12%) and Oceania (n = 1341; 11%). Indonesia (n = 5778; 46%) and Thailand (n = 1483; 12%) were top acquisition countries. DENV-2 (n = 2147; 42%) and DENV-1 (n = 1526; 30%) were predominant serotypes. Conclusion Our analysis highlights Australia’s successful control of locally acquired dengue with Wolbachia. Imported dengue trends reflect both Australian travel destinations and patterns and local epidemiology in endemic countries.


Introduction
Over 3.9 billion people across >128 countries are at risk for dengue. 1 Global incidence has increased 30-fold over the past 50 years 2 and an estimated 100-400 million infections occur annually, with 70% in the Asia-Pacific region. 1,3Dengue results from infection with any of four dengue virus serotypes (DENV-1 to DENV-4), which are transmitted to humans by Aedes aegypti and Aedes albopictus mosquitoes. 2 While many infections are asymptomatic, severe disease may occur.There are no approved antivirals and treatment is supportive. 2 Two dengue vaccines, Dengvaxia and Qdenga, are currently available with varying recommendations and safety implications.
Urbanization and climate change have expanded the geographic range of dengue virus and its mosquito vectors. 4ransmission dynamics vary, with year-round transmission, seasonal and periodic epidemics occurring in different settings. 1,4engue is the leading cause of febrile illness in returned travellers from all continents except Africa, 5 and importation by travellers can drive outbreaks in non-endemic areas with competent vectors. 2 ,6 Australia is not dengue endemic, but importation by viraemic travellers has led to local transmission and periodic outbreaks in areas of northern Queensland with competent vectors. 7 -9andomized and non-randomized field trials in multiple countries over the past decade have demonstrated the effectiveness of the Wolbachia replacement method in significantly reducing dengue transmission, [10][11][12][13][14] when used as an adjunct to existing strategies including vector surveillance and control, mosquito avoidance, disease surveillance, case management, education and emergent vaccination programmes.Field releases of Wolbachia-infected Ae. aegypti as a method for controlling dengue were first conducted in two suburbs of Cairns, in northern Queensland in 2011, 7 ,15 with the aim of stably introducing the maternally inherited intracellular bacterium Wolbachia into the local Ae.aegypti population and thereby reducing the mosquitoes' ability to transmit dengue and other arboviruses. 16Phased deployment of Wolbachia mosquitoes throughout areas of northern Queensland between 2011 and 2019 has significantly reduced local dengue transmission, with sustained high levels of Wolbachia mosquitoes and the effective elimination of autochthonous dengue transmission as a public health concern. 7,15,17n the Australian context, analysis of dengue surveillance data enables both an evaluation of imported dengue trends and the impact of public health control initiatives in areas with local dengue transmission.We report on the epidemiological trends in imported and locally transmitted dengue in Australia from January 2012 to December 2022, and contextualize our findings with broader dengue epidemiological trends in the Asia-Pacific region and Wolbachia mosquito releases in northern Queensland.

Data sources
Dengue cases in Australia.We collated de-identified data on all laboratory-confirmed and clinically diagnosed dengue cases notified to the National Notifiable Disease Surveillance System (NNDSS) between 1 January 2012 and 31 December 2022.Data were provided by the Australian Government's Office of Health Protection in July 2023.Details of variables obtained are in Table S1.

Wolbachia implementation
We gathered data on Wolbachia mosquito release timelines in northern Queensland from published sources 15,17 and operational data provided by the World Mosquito Program.
Traveller movement estimates.We sourced Australian traveller movement data from 2012 to 2022 from the Australian Bureau of Statistics (ABS) (Table S1, available at JTM online).Traveller movements are defined not by individual travellers or trips but rather by international border crossings categorized as short-term (<12 months overseas) resident departures (STRD; 2012-17) and returns (STRR; 2017-22).
Countries were grouped into geographical regions according to the Standard Australian Classification of Countries 2016 (Table S1, available at JTM online).

Dengue cases in other countries of interest.
Where available, we collated publicly accessible data on dengue cases notified to national Departments of Health, other relevant national agencies or to the World Health Organisation for the most common countries of dengue acquisition (Table S2, available at JTM online).We searched PubMed, Google Scholar and Google databases to identify academic papers and reports documenting dengue epidemiology and/or burden of disease data, which were used when other surveillance data were unavailable.

Epidemiological analysis
We estimated the number and annual incidence rate of imported dengue cases in Australia nationally and by jurisdiction.Regionspecific incidences were calculated for the top three dengue acquisition regions and country-specific incidences for the top 10 countries of acquisition.We calculated incidence rates for imported cases by annual traveller movements using the number of imported cases as the numerator and ABS traveller movement data as the denominator.We calculated the annual local incidence of dengue for the top acquisition countries using national case numbers as the numerator and population data for each country as the denominator (Table S1
Country-specific notification incidence rates for cases imported to Australia generally followed quite similar patterns to local dengue incidence, especially for Indonesia, Thailand, Malaysia, Vietnam, Singapore, India, Sri Lanka and Fiji (Figure 5).

Discussion
Our study provides insights into trends in imported and locally acquired dengue in Australia over the last decade, a period marked by record dengue outbreaks in many countries, including Abbreviations: ACT: Australian Capital Territory; NSW: New South Wales; NT: Northern Territory; QLD: Queensland; SA: South Australia; TAS: Tasmania; VIC: Victoria; WA: Western Australia; NI: no information in the Asia Pacific. 18Dengue is among the most frequently imported infections to Australia, 19 and importation into areas of northern Queensland with competent mosquito vectors has historically led to secondary transmission and local outbreaks.We found that imported dengue notifications rose from 2012 to 2016, with peak notification years in 2016 and 2019 aligning with global trends, 18 while locally acquired dengue cases declined to zero concurrent with the implementation of the Wolbachia method in northern Queensland.A sharp decline in imported cases from 2020 to 2021 corresponded with border closures and travel restrictions imposed due to the COVID-19 pandemic.Imported case notifications increased in 2022 following reopening of Australia's borders, but locally acquired cases remained absent.Although DENV serotype distribution varied by region and year, our findings are consistent with a systematic review indicating DENV-2 and DENV-1 as the most common observed serotypes in outbreaks globally from 1990 to 2015. 20hese broad trends are consistent with studies from other non-endemic settings.In the USA, imported cases increased from 2012 to 2016 and were high in 2019. 21In Europe, imported dengue case numbers remained relatively stable between 2015 and 2018 before a significant increase in 2019. 22,23However, Australia's geographic location and the unique travel patterns and migration trends of Australians underpin some notable differences.We found that Asia and Oceania were the most common regions of acquisition, contrasting with the USA, where cases are primarily imported from the Americas. 24,25Similarly, we saw very few cases from Africa, which tends to contribute a higher proportion of imported cases to both the USA and Europe. 22,24These findings highlight the value of Australian dengue surveillance in providing sentinel information on circulating DENV serotypes and outbreaks in acquisition countries, especially in Asia and Oceania.These data can support surveillance programmes in countries with limited resources and surveillance capabilities and assist with risk assessment for travellers to these endemic regions.
Almost two-thirds (64%) of imported cases to Australia originated from Indonesia, Thailand and India, reflecting their popularity as travel destinations and aligning with previous research implicating them as common sources of imported dengue cases globally. 24Importation trends generally aligned with local dengue epidemiology, consistent with previous GeoSentinel reports indicating that dengue trends in returned travellers often mirror local outbreaks in destination countries. 26,27Anomalies in reported dengue incidence in Australia can thus serve as global alerts, signalling possible outbreaks to specific countries and the global health community.
Data regarding specific locations where Australians acquire dengue within Indonesia, where the Wolbachia method has been implemented in the Special Region of Yogyakarta in Java 11 and further expansion is planned, are not available.However, Bali, which reports the highest annual dengue incidence in Indonesia 28 and is a popular travel destination for Australians, is likely an important source of importations. 18Relatively low dengue incidence in Indonesia in 2017-18, potentially combined with  Agung volcano, 29 was reflected in lower imported dengue rates from Indonesia in 2017-18.India experienced large dengue outbreaks in 2021 and 2022, correlating with a marked increase in imported cases to Australia in 2022.This coincided with travel to India rebounding to 77% of 2019 levels.Almost 3% of the Australian population were born in India (Table S1, available at JTM online), making it a common travel destination, particularly for 'visiting friends and relatives' travellers. 30Trends seen for other common acquisition countries, such as the Philippines and Sri Lanka, likely reflect a combination of migration and travel trends, alongside a high local disease burden. 31,32Reasons for some inconsistencies in imported dengue incidence compared to local epidemiology (e.g. for Thailand in 2015) remain uncertain Oceania (n = 1341 cases; 11%) was well represented as a source of dengue acquisition, with Fiji and PNG contributing almost half of all cases from this region.Interestingly, despite travel to Fiji returning to 92% of 2019 levels in 2022, only two imported cases were reported.The reasons for this are uncertain, but a possible contributing factor could be a true decrease in local incidence related to implementation of the Wolbachia method throughout Fiji's three largest cities Suva, Nadi and Lautoka in 2018-19. 33Lack of robust national surveillance data from certain countries in Oceania highlights the value of Australian traveller data in discerning regional dengue trends.
Despite high numbers of imported dengue cases between 2012 and 2019, we observed a substantial reduction in locally acquired dengue cases in Australia over time.This almost certainly reflects the successful large-scale roll-out of Wolbachia mosquitoes in northern Queensland between 2013 and 2019.Long-term monitoring data indicate that Wolbachia has been self-sustaining at a high prevalence in local mosquito populations for a decade post-release. 34The most recent locally acquired dengue outbreak in Queensland occurred in Rockhampton in 2019 (13 laboratory-confirmed cases), 35 an area where Wolbachia mosquitoes were not deployed.This marked the first locally acquired dengue outbreak in Central Queensland in 65 years, emphasizing the need for ongoing monitoring of areas in Australia with potential for local transmission.Given the absence of competent vectors outside QLD, the small number of locally acquired cases (n = 24) notified by other jurisdictions are potential misclassifications.
The success of the Wolbachia method in interrupting local dengue transmission in Australia is an encouraging development for other non-endemic settings with competent mosquito vectors like the USA and Europe, where sporadic outbreaks occur following importation. 21,22,368][39][40][41] For example, the USA has seen a rise in locally acquired cases since 2020, including California's first locally acquired case in October 2023, 42 despite a decrease in imported cases due to COVID-19related travel restrictions. 21Continuing geographic expansion of Ae. aegypti vectors and the dengue virus is predicted, 43,44 presenting substantial challenges in prevention and control efforts.The World Mosquito Program 33 and others 45,46 have undertaken Wolbachia mosquito releases in at least 13 dengue-endemic countries and one non-endemic country (Australia) to date.Further scale-up of the Wolbachia method in Indonesia is underway, and the impact of this initiative on the incidence of dengue among Australian travellers to Indonesia in the years ahead will be of particular interest.The Wolbachia method also has important implications for control of other arboviral infections, 4 as laboratory studies have demonstrated that Wolbachia can modulate replication of yellow fever, Zika and chikungunya viruses in Ae.Aegypti mosquitoes. 47engue vaccines hold promise for reducing the epidemiological and economic burden of dengue in endemic areas, especially when integrated into a multifaceted approach including surveillance and vector control.However, their use in travellers is unclear.At present, two live attenuated tetravalent dengue vaccines, Dengvaxia and Qdenga, are commercially available in a number of countries.For both vaccines, phase III efficacy trials were conducted among children in dengue-endemic settings and data are limited outside of this population. 48Safety concerns around the use of Dengvaxia limit its utility in dengue-naïve individuals, including most travellers, and it has never been commercially available in non-endemic countries. 49For Qdenga, vaccine efficacy differs by serostatus and infecting serotype,  S1) and the literature. 53Data for local dengue cases not available for 2015 for Fiji.Data for local dengue cases not available for 2020 for Fiji.Data for local dengue cases not available for 2021 for Fiji.Data for local dengue cases not available for 2022 for Indonesia, Sri Lanka, Fiji.Data for local dengue cases for Vietnam available until 19th November 2020, 19th December 2021 and 12th December 2022.National dengue data not available for PNG or Timor-Leste.Traveller movement data not available for Timor-Leste.Singapore included given its status as the only other dengue endemic country in the top ten most frequented travel destinations (Appendix, Table S4).
with lower efficacy observed in individuals seronegative for dengue at baseline, especially against DENV-3 or DENV-4. 50n September 2023, the WHO recommended Qdenga be considered in routine immunization programmes in countries with high dengue transmission intensity. 49The benefit in travellers is likely to be for long-term or frequent travel to high-risk destinations in dengue-experienced travellers. 49However, there are potential limitations for dengue-naïve travellers heading to areas where DENV-3 and DENV-4 circulate.While our dataset showed DENV-2 and DENV-1 were predominant among typed cases notified in Australia in 2012-22, DENV-3 and DENV-4 still accounted for a substantial proportion of cases, and DENV-3 has re-emerged in recent large outbreaks in Bangladesh 51 and Brazil. 52DENV serotype data are limited in endemic country national surveillance reports, highlighting a need for enhanced serotype surveillance to guide risk assessment of travellers as well as dengue prevention and control efforts.

Limitations
Limitations of our study include incomplete data on some variables such as region of acquisition and characterization of DENV serotypes.Data captured in the NNDSS are dynamic and subject to retrospective revisions; therefore, the presented data represent a point-in-time analysis of DENV case notifications and may vary from data reported in published NNDSS and jurisdictional reports covering the same period.Notably, other variables of interest such as visitor and immigrant status, pre-travel healthcare, purpose and duration of travel and clinical data are not systematically collected by the NNDSS.This dataset may underestimate the denominator for imported dengue cases as it does not include visitors or migrants entering Australia.In addition, due to the short incubation period of dengue, Australian travellers diagnosed with dengue while overseas might not have been captured in the reported surveillance data.Despite these constraints, our results are based on a large national dataset and provide the most complete record of dengue notification trends over the last decade in Australia.However, since dengue importation is influenced by traveller demographics and destination preference, which can be dynamic and unpredictable, the trends observed in this study may not be generalisable to future patterns.Nonetheless, they highlight the importance of considering country-specific dengue epidemiological trends when assessing travel-associated health risks for travellers.

Conclusion
Our study highlights that notification trends of imported dengue in Australia reflect the travel destinations and patterns of Australian travellers, together with the local epidemiology in endemic countries.Six of the top 10 destinations for Australian travellers in 2022 are highly endemic for dengue, highlighting the importance of ensuring travellers are aware of the risk of dengue and the need to protect themselves from mosquito bites.These findings can aid travel health practitioners to undertake detailed risk assessments and provide accurate education and advice to Australian travellers.
Traditionally, public health measures for dengue prevention and control have encompassed prompt case identification, vector surveillance and control, and human behavioural measures sure as mosquito avoidance.More recently, vaccination programmes have been introduced, with endemic and non-endemic countries employing a mix of approaches depending on their local contexts.The implementation of the Wolbachia method in northern Queensland and other Asia-Pacific and Latin American countries over the past decade has proven its public health value for dengue control.Expanding its implementation in both endemic and nonendemic areas has the potential to achieve sustained control of dengue and other Aedes-borne viruses, benefiting not only local communities but also travellers to these areas.

Figure 1
Figure 1 Number of locally acquired dengue cases and dengue serotype, Australia 2012-22.* Pilot releases of Wolbachia mosquitoes in two suburbs of Cairns occurred in 2011, followed by phased releases across urban areas of northern Queensland in 2013-201915 , 17

Figure 2
Figure 2 Number and notification incidence of imported dengue cases, Australia 2012-22.Abbreviations: STRR: short-term resident return; SE: Southeast.Vertical bars depict the number of cases by region and use the left y-axis, while the lines depict the overall (solid line) and regional (dotted line) notification incidences per 100 000 STRR trips and use the right y-axis.All other regions include Northwest Europe, Southern and Eastern Europe, North Africa and the Middle East, Sub-Saharan Africa, Northeast Asia

Figure 3
Figure 3 Top 10 countries of acquisition by DENV serotype for imported dengue cases, 2012-22.Abbreviations: DENV: Dengue virus.Ranking is based on number of cases imported from each country

Figure 4
Figure 4 Lowest, highest and average incidence (per 100 000 traveller movements) for top countries of acquisition, 2012-19.Abbreviations: STRR: short-term resident return.STRR trips are obtained for each country from ABS data.Data for Timor-Leste not available.Average data for 2012 to 2019 only included as expected decreases in incidence occurred in 2020 following implementation of travel restrictions due to the COVID-19 pandemic

Figure 5
Figure 5 Estimated notification incidence of imported dengue in returned Australian travellers and estimated local incidence of dengue for countries of interest, 2012-22.Abbreviations: STRR: Short term resident return.The solid line represents the incidence in returned travellers and follows the left y-axis.The dashed line represents country specific local incidence and follows the right y-axis.Country-specific dengue case data derived from surveillance data for Indonesia, Thailand, India, Sri Lanka, Malaysia, Vietnam, The Philippines.Local incidence data for Fiji obtained from surveillance reports (Appendix, TableS1) and the literature.53Data for local dengue cases not available for 2015 for Fiji.Data for local dengue cases not available for 2020 for Fiji.Data for local dengue cases not available for 2021 for Fiji.Data for local dengue cases not available for 2022 for Indonesia, Sri Lanka, Fiji.Data for local dengue cases for Vietnam available until 19th November 2020, 19th December 2021 and 12th December 2022.National dengue data not available for PNG or Timor-Leste.Traveller movement data not available for Timor-Leste.Singapore included given its status as the only other dengue endemic country in the top ten most frequented travel destinations (Appendix, TableS4).