Screening patients for potential respiratory infections through nucleic acid sampling was a critical non-pharmacologic intervention during a respiratory infectious disease pandemic. Evaluating the cross-infection risk of crowded activities conducted indoors is necessary. We conducted a field study of this event in a large-scale stadium and measured the variation in CO2 concentrations. We conducted simulations on the stadium's flow field, CO2 and N2O concentration distribution using the CFD method, which was validated by the on-site measured data. We simulated the diffusion of virus aerosols exhaled by the sampled individuals under an all-fresh and primary return air system, respectively. It was found that space morphology and airflow pattern are the main factors influencing the concentration distribution of human exhaled contaminants. Primary return air systems can reduce the infection risk for sampled individuals by the filtration system and improve the airflow pattern. With the ratio of fresh air to return air of 1.2, the primary return air systems with 70 %, 85 %, and 95 % filtration levels can reduce the infection risk by an average of 54.5 %, 58.3 %, and 62.1 %, respectively, compared to all-fresh air system. It is necessary to optimize the processes and queue setups for nucleic acid sampling activities in the room based on the room airflow characteristics and the return air system using an efficient filtration system.