Laura Mulas 1, Maxime Beaulieu 1, Antonin Bal 2, Florence Débarre 3, François Blanquart 4, Jérémie Guedj 1
1 IAME Inserm, Université Paris Cité (F-75018 Paris, France), 2 HCL and CIRI, Inserm U1111, CNRS UMR5308, ENS de Lyon, UCBL Lyon 1 (Lyon, France), 3 Institute of Ecology and Environmental Sciences Paris (IEES Paris), Sorbonne Université, CNRS, IRD, INRAE (Paris, France), 4 Centre interdisciplinaire de recherche en biologie, Collège de France (Paris, France)
Introduction:
Influenza viruses, respiratory syncytial virus (RSV), and SARS-CoV-2 drive recurrent seasonal epidemics and impose a substantial healthcare burden[1–3]. Since the COVID-19 pandemic, molecular PCR testing has become routine in community laboratories, enabling large-scale detection of respiratory pathogens. Beyond diagnosis, these assays generate large-scale quantitative cycle threshold (Ct) data that provide indirect measures of viral load and remain underexploited for mechanistic modelling of infection dynamics.
In France, the RELAB network, coordinated by the Centre National de Référence des Virus des Infections Respiratoires (CNR, Hospices Civils de Lyon and Institut Pasteur, Paris), aggregates diagnostic data from a large consortium of community laboratories performing routine triplex PCR assays for SARS-CoV-2, Influenza viruses, and RSV.
While PCR-based viral dynamics have been studied for SARS-CoV-2[4,5], comparative mechanistic analyses across other major respiratory viruses remain limited. In addition, vaccination coverage and effectiveness vary substantially across seasons and viral subtypes. Leveraging routinely collected community PCR data available in near real time therefore offers a unique opportunity to quantitatively assess how vaccination modifies within-host viral trajectories.
Objectives:
This study aimed to develop a mechanistic population modelling framework to characterise viral load dynamics in symptomatic SARS-CoV-2, Influenza and RSV infections using community-based PCR data from the RELAB network. Specifically, we sought to quantify the contribution of host characteristics, epidemic context, and vaccination to inter-individual variability in virus-specific viral trajectories.
Methods:
To ensure comparability across laboratories and PCR platforms, Ct values were harmonised using laboratory-specific normalisation rules. This procedure reduced measurement heterogeneity before model fitting.
We then developed a simplified mechanistic target cell-limited model to describe viral kinetics[6,7]. Given the cross-sectional nature of community data, the model was parameterised to focus on identifiable composite parameters governing viral proliferation and clearance.
Parameters were estimated using nonlinear mixed-effects modelling with the SAEM algorithm implemented in Monolix. Covariate effects (age, sex, epidemic period, and vaccination status) were evaluated on key mechanistic parameters governing viral expansion and clearance for each virus using an automated stepwise covariate modelling procedure (COSSAC).
Results:
Analysis of community-based PCR data from the 2024-2025 and 2025-2026 winter seasons revealed substantial heterogeneity in viral load dynamics across viruses and epidemic periods. Seasonal differences were particularly pronounced for Influenza A and RSV, likely reflecting the circulation of distinct viral variants within the study period.
Age significantly influenced viral clearance in a virus-specific manner. Children cleared SARS-CoV-2 infections faster than adults, whereas Influenza infections were associated with slower clearance in children, consistent with previous reports describing distinct age-dependent viral dynamics across respiratory viruses[8,9]. Sex-related differences were also observed, with women exhibiting faster viral clearance than men for Influenza infections, in line with prior epidemiological and clinical studies[10].
Vaccine effectiveness estimated from community data was substantial only for Influenza B during the study period. In vaccinated individuals, the model estimated a reduction of approximately two days in time to viral clearance and an approximate fourfold decrease in peak viral load. No significant vaccination effect on viral dynamics was detected for SARS-CoV-2 or Influenza A, which may reflect the lower vaccine effectiveness reported for these pathogens during the study period[11].
Conclusion:
This study provides a unified mechanistic population modelling framework to characterise viral load dynamics from large-scale community PCR data and to quantify how host, seasonal, and vaccination factors shape infection trajectories.
However, the limited availability of observations during the late phase of infection may bias the estimation of trajectory-derived quantities such as time to viral clearance, which appears to be overestimated compared with values reported in the literature[12,13].
By integrating routinely collected diagnostic data, this framework contributes to real-time epidemiological surveillance of respiratory viruses. As multiplex PCR testing expands, the availability of harmonised measurements across pathogens may enable future comparative analyses of viral dynamics and provide opportunities to investigate inter-virus interactions and co-infections, ultimately helping bridge within-host viral dynamics and population-level respiratory surveillance.
References:
[1] Acute respiratory infections: the forgotten pandemic. Bull. World Health Organ. 76, 101–107 (1998).
[2] Vos, T. et al. Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet 396, 1204–1222 (2020).
[3] Cui, C et al. Disease burden and high-risk populations for complications in patients with acute respiratory infections: a scoping review. Front. Med. 11, 1325236 (2024).
[4] Beaulieu, M. et al. The analysis of a large dataset of PCR-test results in the community reveals the impact of age, variants and vaccination status on SARS-CoV-2 viral dynamics (2025).
[5] Hay, J. A. et al. Estimating epidemiologic dynamics from cross-sectional viral load distributions. Science 373, eabh0635 (2021).
[6] Jeong, Y. D. et al. Designing isolation guidelines for COVID-19 patients with rapid antigen tests. Nat. Commun. 13, 4910 (2022).
[7] Kim, J. et al. Comparing Cross-Sectional and Longitudinal Study Designs for Accurate Viral Dynamics Estimation: Insights From the NBA Cohort Data. J. Med. Virol. 98, e70823 (2026).
[8] Fryer, H. R. et al. Viral burden is associated with age, vaccination, and viral variant in a population-representative study of SARS-CoV-2 that accounts for time-since-infection-related sampling bias. PLOS Pathog (2023).
[9] Cauchemez, S. et al. Closure of schools during an influenza pandemic. Lancet Infect. Dis. 9, 473–481 (2009).
[10] Klein, S. L. & Flanagan, K. L. Sex differences in immune responses. Nat. Rev. Immunol. 16, 626–638 (2016).
[11] Blanquart, F. Influenza vaccine effectiveness against detected infection in the community, France, October 2024 to February 2025.
[12] Zhang, S. & Agyeman, A. A. SARS-CoV-2 viral dynamic modeling to inform modelselection and timing and efficacy of antiviral therapy.
[13] Canini, L. & Carrat, F. Population Modeling of Influenza A/H1N1 Virus Kinetics and Symptom Dynamics. JOURNAL OF VIROLOGY (2011).
Reference: PAGE 34 (2026) Abstr 11861 [www.page-meeting.org/?abstract=11861]
Poster: Drug/Disease Modelling - Infection