Assessing the burden of respiratory syncytial virus disease in Australia

National reporting for all age groups is needed to accurately determine the full burden of disease

Respiratory syncytial virus (RSV) is a pathogen of global importance, responsible each year for more than 33 million infections, 3.2 million hospitalisations, and 59 600 in‐hospital deaths of children under 5 years of age.1 The full burden of disease is likely to be much higher, as 50% of deaths occur outside hospitals.2 The World Health Organization regards RSV vaccine development as a global priority,3 and RSV vaccines have been short‐listed for the Global Vaccine Alliance (Gavi) vaccine investment strategy; several vaccine candidates and monoclonal antibodies are being investigated, and it is anticipated that commercial vaccines will be available by 2025.4,5 Gaps in knowledge about the epidemiology of RSV — accurate burden of disease estimates, its long term effects, seasonal transmission dynamics, the cost of RSV‐related disease — need to be filled to guide vaccine policy and implementation.6

Interim results of a phase 3 trial of a maternal RSV vaccine were published by Novavax in March 2019.7 While efficacy with respect to the primary objective of preventing medically significant RSV lower respiratory tract infections in infants during their first 90 days of life (39%; 97.5% confidence interval [CI], –1% to 64%) was lower than the pre‐specified success criterion of 41%, it prevented hospitalisations for laboratory‐confirmed RSV acute lower respiratory infections with 44% efficacy (95% CI, 20–62%).7

We must quantify the burden of RSV in Australia across all age and risk groups to determine whether maternal vaccination would achieve significant improvements in health outcomes. In this edition of the MJA, Saravanos and her colleagues have attempted to reduce some of the gaps in our epidemiological knowledge by describing RSV‐associated hospitalisations for all age groups on the basis of national hospitalisation data.8 Their article raises several critical questions.

The first and major limitation of their study concerns the reliability of the International Classification of Diseases, version 10, Australian modification (ICD‐10‐AM) diagnosis codes with which the authors identified RSV‐related hospitalisations. Without confirmation by laboratory data, hospital diagnosis codes are likely to underestimate the true burden of disease caused by RSV. Further, the rationale for selecting the specific ICD‐10‐AM codes used by the authors (J12.1, J20.5, J21.0, B97.4) was not clear. RSV is not a notifiable disease, but patients are often tested for it when tested for other respiratory viruses, including influenza, for which there are numerous sources of national surveillance data.

The analysis by Saravanos and colleagues should motivate considering national reporting of RSV infections in all age groups. Western Australia is unique in its ability to use record linkage for combining ICD‐coded hospitalisation data with routine laboratory data. This has allowed us to identify laboratory‐confirmed RSV hospitalisations of children with various ICD‐10‐AM codes, including the nonspecific codes B34.9, J06.9, J22, and J45.9.9 Our findings indicate that the number of RSV‐related hospitalisations, particularly of children, was probably underestimated by Saravanos and colleagues because of their limited range of ICD‐10‐AM codes. Routine laboratory data are available in most Australian jurisdictions and more linkages of clinical and laboratory datasets should be undertaken, as this approach may elucidate reasons for the large variations in reported rates, which may reflect differences between states in testing practices. Additionally, RSV has emerged as a leading cause of pneumonia;10 including only one pneumonia ICD‐10‐AM diagnosis code in the analysis is likely to have led to Saravanos and colleagues further underestimating the RSV burden.

A second problem is the difference in burden between Aboriginal and Torres Strait Islander Australians and other Australians. The reliability of Indigenous status data in the National Hospital Morbidity Database is unknown. Algorithms have been developed for improving the quality of Indigenous identification in administrative data, ideally by including a combination of datasets.11 As with influenza and most other infectious diseases, the burden of RSV in Indigenous Australians is probably higher than among other Australians, but the magnitude of the difference is likely to be underestimated until more accurate data on Indigenous identification are available.

Saravanos and colleagues provide novel and important data on the high rates of RSV‐associated hospitalisations of older adults, an age group in which rates of routine testing for respiratory viruses are low but rising, particularly as diagnostic testing for RSV is progressively combined with that for influenza viruses. We need to better understand RSV mortality in older people, as estimates are likely to be influenced by testing rates. Saravanos and her co‐authors also report an important number of deaths in this age group in cases that were coded as RSV pneumonia. Indirect and herd immunity effects of preventing RSV transmission by infants and young children should be a further research focus.

RSV remains a major cause of morbidity for Australians, and we need to focus on developing routine surveillance mechanisms, consider adding RSV to the national notifiable diseases list, and encourage linkage of laboratory and clinical data, critical for determining hospitalisation numbers attributable to RSV and the burden missed by current testing practices. Similar efforts are being undertaken by the Respiratory Syncytial virus Consortium in Europe (RESCEU) network (resc-eu.org).