Urban environment and RSV: a retrospective observational study of neighbourhood factors associated with the risk of severe disease in the infant population of a metropolitan area, Lyon, France

Background: Severe acute respiratory infections (SARI) caused by a human respiratory syncytial virus (RSV) are a leading cause of hospitalisation among infants. We aimed to estimate the incidence of RSV SARI across neighbourhoods in the Lyon metropolitan area and to assess how urban environmental factors at the metropolitan scale are associated with spatial variation in incidence. Methods: Laboratory-confirmed cases of RSV SARI (< 2 years of age) were extracted from the university hospitals of Lyon laboratory database 2015–2023. We calculated and mapped the incidence to assess spatial variation. Remote sensing data were used to derive spectral indices characterising the metropolitan area and to model temperature and air humidity at this scale. Temporal and spatial regression models were fitted using variables selected based on prior knowledge and data availability. Results: Cumulative incidence varied significantly across neighbourhoods (0 to 1166 cases per 100,000 people at risk; p-value < 0.001). The best spatial multivariate model (r² = 0.39, Akaike information criterion (AIC) = 6103), explained a substantial portion of this variation, and included neighbourhood median income that was negatively associated with incidence; and neighbourhood winter temperature as well as particulate matter < 10 µm pollution, both positively associated with incidence. Additionally, urban index (UI) and normalised difference moisture index (NDMI) demonstrated strong (p-value < 0.001) univariate associations (UI: r² = 0.23, AIC = 6216; NDMI: r² = 0.21, AIC = 6229), accounting for a significant portion of the incidence variation. Conclusions: Substantial neighbourhood-level differences in RSV SARI incidence exist in a large European metropolis. These differences are associated with the urban environment such as particulate pollution. The use of spectral indices shows promise in identifying vulnerable populations within cities to guide public health measures and to integrate public health and urban planning.