Viral Dispersion in the ICU: The Wind Effect

To the Editor:

We read with great interest the article by Avari and colleagues on the topic of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) dispersion associated with respiratory support devices in the ICU (1). In the article, the authors focused their attention on medical devices that influence the emission of aerosol production by a contagious patient. If the contribution of Avari and colleagues is of most importance, a whole area of physical conditions is not specified and could have a major impact on the spreading of the virus. Indeed, if respiratory support could be aerosol generating procedures, airflow in the room because of the ventilation system or an open window could change the dispersion of SARS-CoV-2 in the environment of the patient. In many ICUs, to avoid healthcare workers’ contamination owing to SARS-CoV-2, windows are open in the rooms of patients with coronavirus disease (COVID-19) to improve air renewal.

To identify the risk of environmental contamination, we compared PCR technology for the qualitative detection of nucleic acids from SARS-CoV-2 on surface samples to scale resolved time accurate computational fluid dynamics simulations solver based on the Lattice-Boltzmann Method (PowerFLOW). We simulated consequences of the wind on environmental contamination and ventilation systems while a detailed human model was emitting both airflow and aerosols centered around 2 μm in a detailed three-dimensional (3D) digital twin of the rooms of our ICU. Samples were performed in two different rooms on the same floor of the same building during periods with various winds.

We built a 3D digital twin of the patient’s room, including a detailed ventilation system and medical devices. This model was built using two-dimensional blueprints of the hospital, pictures of the rooms, and ventilation system settings as set during the sampling period. Exterior wind conditions were directly integrated to the simulation model and a detailed human model was emitting both airflow and aerosols centered around 2 μm.

The combination of computational fluid dynamics and PCR test demonstrated that the contamination of the surfaces is directly correlated to the airflow in the room. Surprisingly, for the same room, the same contaminated patient, and the same sampling method, the samples switched from all negative to almost all positive between April 2 and April 6, 2020. During this period of 4 days, the wind switched from an east to a northwest direction while the temperature was showing a small rise of 2°C and relative humidity was stable (60%). By simulating the exact same condition and considering the location of the room in the hospital, it appears that external air now entered through the window, mainly because of the wind direction. This new airflow created a strong recirculation inside the room, changing the way aerosols were transported and deposed. Most of the surfaces that showed no deposit during the first simulation appear now to be impacted by droplets deposition. These newly contaminated surfaces in the simulation match with positive PCR test points during this period.

The presence of SARS-CoV-2 on surfaces is also due to hand carrying from patient or healthcare providers. In our study, to highlight airborne transportation, all surfaces were cleaned 8 hours before sampling the surfaces while the patient did not move from the bed.

The present work tackles different airflow-dictated issues such as propagation of aerosols emitted by a patient with COVID-19, the impact of the ventilation system, and the impact of external influences such as wind direction and intensity. Finally, simulations can accurately predict SARS-CoV-2–contaminated surfaces.

Based on our study, we encourage physicians to focus not only on the patient with COVID-19 while approaching viral spreading but also on factors independent of the patient, such as ventilation system and exterior wind condition when opening windows.