Evidence for an increase in the intensity of inter‐seasonal influenza, Queensland, Australia, 2009‐2019

Abstract Background Inter‐seasonal influenza cases have been increasing in Australia. Studies of influenza seasonality typically focus on seasonal transmission in temperate regions, leaving our understanding of inter‐seasonal epidemiology limited. We aimed to improve understanding of influenza epidemiology during inter‐seasonal periods across climate zones, and explored influenza intensity and strain dominance patterns over time. Methods Queensland state‐wide laboratory‐confirmed influenza notifications and public laboratory influenza test data from 2009‐2019 were described by demographics, time period, region and strain type. We compared influenza intensity over time using the WHO Average Curve method to provide thresholds for seasonal and inter‐seasonal periods. Results Among the 243 830 influenza notifications and 490 772 laboratory tests reported in Queensland between 2009 and 2019, 15% of notifications and 40% of tests occurred during inter‐seasonal periods, with 6.3% of inter‐seasonal tests positive. Inter‐seasonal notifications and tests substantially increased over time and increases in weekly proportions positive and intensity classifications suggested gradual increases in virus activity. Tropical inter‐seasonal activity was higher with periods of marked increase. Influenza A was dominant, although influenza B represented up to 72% and 42% of notifications during some seasonal and inter‐seasonal periods, respectively. Conclusions Using notification and testing data, we have demonstrated a gradual increase in inter‐seasonal influenza over time. Our findings suggest this increase results from an interplay between testing, activity and intensity, and strain circulation. Seasonal intensity and strain circulation appeared to modify subsequent period intensity. Routine year‐round surveillance data would provide a better understanding of influenza epidemiology during this infrequently studied inter‐seasonal time period.


| BACKG ROU N D
In temperate regions human influenza typically has a strong seasonal cycle where influenza activity increases between late autumn and early spring, before returning to baseline activity during warmer months. 1,2 Influenza cases during the summer inter-seasonal period are often considered to be sporadic imported cases unlikely to cause substantial ongoing transmission. [3][4][5] However, recent evidence suggests that inter-seasonal transmission patterns are more complex than previously thought. 3 In tropical and sub-tropical regions, inter-seasonal influenza patterns are variable with some areas experiencing increased activity during rainy seasons, and others not experiencing a well-defined season. 2,6 Regardless of the climate zone, outbreaks of influenza may also occur outside of the typical season as evidenced by the large and widespread influenza outbreaks across Australia during the 2018/19 inter-seasonal period. 7 Changes to influenza reporting, testing capacity and testing behaviour likely play a role in increased recognition of influenza infections during inter-seasonal periods, both globally and in Australia. 1 Nonetheless, higher than average inter-seasonal influenza notifications, beyond what could be accounted for by increased testing and reporting, have been observed previously in Australia during the 2010/11 and 2018/19 inter-seasonal periods. 7,8 Recent Australian analyses provide insights into Australian inter-seasonal activity, transmission and testing practices, yet previous studies of influenza seasonality typically focus on transmission within seasonal periods in temperate regions. 1,3,7,9 Our understanding of the epidemiology of inter-seasonal periods across different climatic regions over time therefore remains limited.
Here we use routinely collected surveillance data from 2009 to 2019 to summarize the epidemiology of influenza during inter-seasonal periods in Queensland over a 10-year period and explore influenza intensity and strain dominance patterns over time. We focus our analyses on the Australian state of Queensland, which by spanning temperate, sub-tropical and tropical climates provides insights into patterns of seasonality across different climatic regions.

| Study setting and data sources
Queensland is a state with a population of approximately 5 million people located in Australia's northeast. It has a varied climate including temperate, sub-tropical and tropical zones. Queensland has well-established influenza surveillance systems which monitor activity using laboratory-confirmed influenza notifications, public laboratory influenza testing data and hospital admissions for influenza to Queensland public hospitals. Laboratory-confirmed influenza is a nationally notifiable condition in Australia and has been notifiable on pathological diagnosis in Queensland since 2001. Confirmed cases are notified on laboratory definitive evidence; virus isolation by culture, detection by nucleic acid testing, or antigen detection in appropriate respiratory specimens or seroconversion or a fourfold or greater rise in an antibody titre to influenza virus. Notification is also made on the detection of a single high titre of IgA antibody.
Notifications of laboratory-confirmed influenza from both public and private laboratories are recorded in Queensland's Notifiable Condition System (NoCS). The Queensland public laboratory information system (AUSLAB), holds laboratory test request and result records of all public hospital inpatients and outpatients, as well as testing records for community clinics and prisons in Queensland.
Routinely collected data from the beginning of the 2009 season (01 May) to the end of the 2019 season (03 November) were used for this analysis. Laboratory-confirmed influenza notifications and data for all individuals tested for influenza using PCR were extracted from NoCS and AUSLAB, respectively. As serology was infrequent during this study period, representing only 2.45% of total testing and 1.90% of positive tests, only notifications and testing data from PCR results, including GeneXpert data, were used.

| Analysis
Notification and testing data were analysed as both overall period (seasonal/inter-seasonal), individual season/inter-seasonal period, and weekly totals. Weeks were defined using International Organisation for Standardisation (ISO) 8601 standard weeks.

| Intensity and strain dominance
To characterize and compare influenza intensity over time, thresholds were set for seasons and inter-seasonal periods separately using the World Health Organization (WHO) average curve method. 11 Historic data from 31 May 2010 (beginning of the 2010 season) to 03 November 2019 were aligned on the median week of peak activity and assigned thresholds of "no activity" (below annual median value), "low" (between the annual median value and the upper 40% confidence interval (CI) of the mean peak value of the average curve), "moderate" (between the upper limit of the 40% CI and 90% CI of the mean peak value of the average curve), "high" (between the upper limit of the 90% and 97.5% CI of the mean peak value of the average curve), and "extraordinary" (above the upper limit of the 97.5% CI of the mean peak value of the average curve). Due to large variance in peak values, CIs were calculated using a geometric mean. Such composite influenza measurements are considered a better proxy indicator of influenza incidence than either notification or laboratory data alone as they improve representativeness and account for testing practices and behaviours. 12,13 Strain dominance was defined as the largest proportion of strain type among all notifications during that period. Influenza B co-circulation was defined as where influenza B accounted for ≥20% of all notifications for the period.

| Inter-seasonal epidemiology
Increases in notifications and influenza tests over the study period were most evident among the ≥65-year age group; notifications and tests increased each inter-seasonal period on average by 61% and 59%, respectively. Thirty-four per cent of all tests during interseasonal periods were for children younger than 5-years of age and testing in this age group was high across all inter-seasonal periods, ranging between 5410 tests (2009/10) and 9539 (2015/16).
However, the proportion positive for children younger than 5 years of age was the lowest of all age groups (1592/68 144, 2.3%) and increases in notifications over the study period were consistent across all age groups, apart from the ≥65-year age group.
The The overall ratio for inter-seasonal to seasonal periods was 0.19 for notifications and 0.74 for testing. When assessed with other categories, higher testing ratios, comparing inter-seasonal to seasonal periods, were observed among children younger than 5-years of age (0.93), males (0.77), and the tropical region (0.94) ( Table 1).
The proportion positive ratio was also highest for the Tropical region

| Intensity and strain dominance
Using the intensity threshold levels, the majority of inter-seasonal periods were categorized as either "no activity"
Therefore, countries and regions, specifically those with tropical or varied climates, should look to undertake routine year-round influenza surveillance stratified to account for climatic zones and strain circulation. With potential interconnectedness between seasonal and inter-seasonal influenza activity, surveillance of inter-seasonal influenza may also have value in contextualizing or anticipating seasonality and intensity over time. Further research into the potential flow on impacts of seasonal intensity and strain circulation on inter-seasonal and seasonal periods over a longer timeframe may provide greater insights into drivers of influenza seasonality.

ACK N OWLED G EM ENTS
The authors thank Nicole Burt and Mohana Rajmokan for their assistance extracting data used in this study.

CO N FLI C T O F I NTE R E S T
None declared.

PATI ENT CO N S ENT S TATEM ENT
All data used by this study was from state-wide routine surveillance of a notifiable condition and were de-identified.

PE R M I SS I O N TO R E PRO D U CE M ATE R I A L FRO M OTH E R S O U RCE S
No permissions required.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are subject to third part restrictions. The data that support the findings of this study are available from Queensland Health.
Restrictions apply to the availability of these data, which were