Abstract
Objectives: Respiratory viral infections are a major global health burden, often involving co-circulating pathogens. In 20-30% of cases, multiple viruses are detected simultaneously, with certain viral pairings attenuating disease while others exacerbate it. The biological mechanisms underlying these contrasting outcomes remain incompletely understood. To address this gap, we employed mathematical models to study coinfections involving influenza A virus (IAV) with respiratory syncytial virus (RSV), rhinovirus (RV), and SARS-CoV-2 (CoV2). We aimed to identify possible mechanisms of viral interference or synergy and assess how infection sequence, timing, and viral pairing influence within-host dynamics and disease severity.
Methods: We extended a basic viral kinetics model based on ordinary differential equations under two hypotheses: (1) viruses compete for a shared pool of target cells; (2) viruses infect distinct epithelial cell populations, reflecting tissue tropism. We calibrated these models to viral load data from ferrets and humans with mono- and coinfections. Parameters were estimated using nonlinear mixed-effects modeling and the SAEM algorithm in Monolix 2019R1. We incorporated weight-loss data from murine coinfection studies, on IAV-RV, RV-IAV, and IAV-CoV2, to interpret clinical relevance. Interaction functions were included to simulate viral enhancement or inhibition at various infection stages (i.e., infection, production, clearance). Total viral burden was quantified by computing the area under the viral load curve using Python’s scipy.integrate.trapz.
Results: Modeling revealed that IAV reduced RSV production rates, whereas RSV delayed IAV-infected cell clearance. Simulations of various infection orders and timings suggested that the heightened severity observed in murine IAV-RV and IAV-CoV2 coinfections stems from delayed clearance of IAV-infected cells. Conversely, when IAV followed RV, disease severity was reduced - a phenomenon our model replicated by lowering RV clearance in the presence of IAV.
Conclusions: Mathematical modeling of respiratory viral coinfections can generate testable hypotheses and clarify how timing, order, and specific viral interactions influence disease outcomes. Our findings highlight the importance of infection sequence and support future efforts to inform targeted antiviral therapies and public health strategies.