Respiratory syncytial virus-attributable hospitalizations among adults in high- and middle-income countries: application of the Global Burden of Disease framework.

Abstract

Background: Respiratory syncytial virus (RSV) in adults is typically underdiagnosed due to non-specific symptoms, infrequent routine testing, and low-test sensitivity; consequently, its impact is not well understood. To address this gap, we developed a novel approach to estimate adult RSV-related hospitalizations, leveraging methods from the Global Burden of Disease (GBD) study.

Methods: We collected aggregated clinical data from hospital statistics and insurance claims on respiratory and cardiorespiratory hospitalizations and RSV activity proxies for age groups 18-59 years, 60-74 years, ≥60 years, and ≥75 years in 15 countries (Argentina, Brazil, Canada, Chile, Georgia, Germany, Greece, Ireland, Italy, Japan, Mexico, New Zealand, Poland, Spain, and the United States) between 1992 and 2021. In addition, we collected RSV surveillance data, i.e., the percentage of samples tested positive for RSV from the WHO GISRS platform-the Global Influenza Surveillance and Response System and from country-specific reporting platforms for countries from North and South America, Europe and Asia, covering the years 2015-2023. Using the GBD comparative risk assessment framework, we estimated exposure-response relationships between RSV activity and hospitalizations using generalized additive models (GAMs), adjusting for trend, seasonality, meteorological influence and influenza activity, between the years 2015-2019, and calculated the population attributable fraction (PAF) and RSV-attributable hospitalizations. We evaluated the predictive power of surveillance-based versus hospital-based RSV proxies based on adjusted R², and generalized cross-validation (GCV) score.

Findings: We identified significant relationships (p-value < 0.01) between RSV activity and increased respiratory and cardiorespiratory hospitalizations among adults. Generally, hospital-based RSV proxies predicted hospitalization better than surveillance-based proxies. RSV-attributable hospitalization rates and PAFs varied substantially by age and country. The highest annual RSV-attributable hospitalization rates were estimated for individuals 75 years and older, ranging from 110.9 (95% uncertainty interval [UI]: 66.9-156.1, median: 113.5, inter quartile range [IQR]: 10.4) per 100,000 population in Argentina for respiratory hospitalizations to 1199.8 (1087.0-1313.8, 1209.5, 88.9) per 100,000 in New Zealand for cardiorespiratory hospitalizations. The lowest RSV-attributable hospitalizations, for respiratory and cardiorespiratory diseases, were found for adults aged 18-59 years in Spain with 5.0 (95% UI: 0.8-9.3) hospitalizations per 100,000 for the hospital-based proxy.

Interpretation: Innovations introduced by this analysis include non-parametric modelling of the exposure-response relationship between RSV activity and hospitalizations and evaluating the predictive reliability of two RSV proxies. Our findings highlight the substantial adult RSV disease burden, provide estimates for countries with no prior data (particularly those in (sub)tropical climates such as Mexico and Brazil), and illustrate the considerable geographic variability in adult RSV incidence. These results can guide future research, interventions, and policy decisions, including those involving adult RSV vaccines.

Funding: This study was sponsored by Pfizer Inc.