key: cord-0985632-k9tdsrq4 authors: Dudalski, Nicholas; Mohamed, Ahmed; Mubareka, Samira; Bi, Ran; Zhang, Chao; Savory, Eric title: Experimental investigation of far‐field human cough airflows from healthy and influenza‐infected subjects date: 2020-05-04 journal: Indoor Air DOI: 10.1111/ina.12680 sha: cc26e2a787331f61ea442b084184725fa4d9ba62 doc_id: 985632 cord_uid: k9tdsrq4 Seasonal influenza epidemics have been responsible for causing increased economic expenditures and many deaths worldwide. Evidence exists to support the claim that the virus can be spread through the air, but the relative significance of airborne transmission has not been well defined. Particle image velocimetry (PIV) and hot‐wire anemometry (HWA) measurements were conducted at 1 m away from the mouth of human subjects to develop a model for cough flow behavior at greater distances from the mouth than were studied previously. Biological aerosol sampling was conducted to assess the risk of exposure to airborne viruses. Throughout the investigation, 77 experiments were conducted from 58 different subjects. From these subjects, 21 presented with influenza‐like illness. Of these, 12 subjects had laboratory‐confirmed respiratory infections. A model was developed for the cough centerline velocity magnitude time history. The experimental results were also used to validate computational fluid dynamics (CFD) models. The peak velocity observed at the cough jet center, averaged across all trials, was 1.2 m/s, and an average jet spread angle of θ = 24° was measured, similar to that of a steady free jet. No differences were observed in the velocity or turbulence characteristics between coughs from sick, convalescent, or healthy participants. large droplets play a significant role in virus transmission, they are expected to fall more quickly and are less likely to be inhaled. Several reviews support the claim that droplet nuclei smaller than 5 μm behave much like a gas and are capable of remaining suspended within the air for long periods of time, and may contribute to airborne virus transmission. 7, 8 Several studies have tried to determine the size distribution of aerosolized particles produced by coughing, but the distributions varied between experiments. [9] [10] [11] [12] [13] It has been indicated that 99% of all expired particles are smaller than 10 μm and they can be easily inhaled by a susceptible host. By taking measurements of particles in their equilibrium states (after all volatile water content has evaporated), a multi-modal distribution of particle sizes with peaks at 1 and 100 μm has been observed. 11, 12 Biological sampling has confirmed the presence of viable pathogens within these aerosols. [14] [15] [16] Experimental investigations concerning subjects who have been naturally infected with influenza have shown that 65% of influenza RNA was found in particles smaller than 4 μm, although the presence of influenza RNA does not necessarily indicate the presence of viable viruses. 15 However, such investigations were conducted at or near the mouth of subjects. The velocity at which infectious particles are introduced into the environment influences the potential for transmission. Several investigations have been conducted to examine the flow behavior of coughs at or near the mouth of human subjects, [17] [18] [19] [20] [21] [22] [23] [24] [25] but only one investigation has examined cough flow behavior beyond this region. 26 While determining the cough flow behavior close to the mouth can be useful for determining boundary conditions for numerical or physical simulations, the flow behavior in the far field is more important when assessing the potential for virus transmission, since such separation distances are more common for most human interactions. Volumetric flow rates, volumes, and peak velocity times of coughs have been determined using spirometry techniques, 17, 21 whereas shadowgraph techniques, 22 video imaging, 24, 25 and particle image velocimetry (PIV) [18] [19] [20] 23, 26 have been used to visualize the flow field. The experimental results vary significantly between experiments and between subjects, and no well-defined model exists for near-field or far-field cough behavior. Furthermore, the separation distance that distinguishes between the near field and far field of cough flow behavior has not been defined. All previous investigations of the flow field have been conducted with healthy subjects and so there is no evidence indicating that coughs from subjects who have been naturally infected with respiratory viruses behave like those from healthy subjects. Statistical issues also become evident, due to the small number of subjects recruited for the experiments. Very little is known about the production and dispersion of viral bioaerosols, even though such information is critical in healthcare settings during viral outbreaks. Viable influenza particles have been recovered through air sampling at hospitals, at health centers, and on airplanes, [27] [28] [29] and several factors have been shown to influence airborne transmission of the influenza virus. Guinea pig models have demonstrated that low relative humidity (RH) and low temperatures enhanced the transmission of the influenza virus, whereas high RH and high temperatures interrupted transmission. [30] [31] [32] [33] Despite this, there is a widespread adoption of the "3 ft/1 m" and "6 ft/2 m" rules, 34, 35 which have considered such separation distances from patients infected with respiratory viruses to be safe, without any evidence to support the claim. The objective of the present investigation was to test the "3 ft/1 m rule," by conducting velocity measurements and bioaerosol sampling at x = 1.0 m from the mouth of human subjects. Based on an average mouth opening diameter of D = 0.02 m, 21 this region is calculated to be approximately x = 50D. Experiments were conducted to map the flow field of human coughs at greater distances than those studied previously and to develop a model for transient cough jet behavior. Subjects who have been naturally infected with influenza participated in experiments while they were sick and again when they had recovered. A cohort of healthy volunteers was recruited as a reference to assess any difference in the aerodynamic behavior of the coughs from the two groups. The velocity measurements were also used to validate computational fluid dynamics (CFD) models based on unsteady Reynolds-averaged Navier-Stokes (URANS) and large eddy simulation (LES) methods developed by Bi. 36 While it has been demonstrated that viral aerosols are typically expired from breathing, 37 the present investigation focuses solely on the fluid mechanics of coughing, since there is a potential for the transportation of aerosols and droplets further from the source. The experimental chamber 24 consisted of a 1.81 m × 1.81 m × 1.78 m enclosure, with all interior surfaces painted black, except for a glass window allowing optical access and a glass panel in the center of the chamber ( Figure 1 ). The dimensions of the chamber were selected so that the cough airflow was not noticeably influenced by the chamber walls. A pear-shaped opening, with a padded head rest and chin rest, fixed the participant's head in place, but allowed the participant to cough into the chamber with their nose and mouth unobstructed. Subjects were asked to cough 3 times each in separate particle image velocimetry (PIV) and hot-wire anemometry (HWA) experiments with aerosol sampling. Participants waited for approximately • The present work can be useful when quantifying air movement from coughs in healthcare settings. • The results of the study can be used to identify safe separation distances and to develop preventative measures for the mitigation of person-to-person virus transmission. 2 minutes between coughs to minimize the residual air motion within the chamber. PIV has commonly been used as a non-intrusive technique to measure airflow fields due to its accuracy and spatial resolution. For PIV experiments, the chamber was seeded with aerosolized titanium dioxide (TiO 2 ) particles (ranging between 0.15 and 0.47 µm with 69% of particles between 0.34 and 0.43 µm) that were dried in a vacuum oven. The particles were stored in a drum, which was placed on top of a loudspeaker that vibrated to aerosolize the powder. The seeding particles were carried into the chamber from the drum by a 30 kPa airline. Particles were illuminated by a 120 mJ per pulse, 532-nm Nd:YAG laser, which was directed through an angled mirror and a cy- The time histories were extracted at the cough jet center, as well as at the chamber centerline and the HWA location, x = 1 m away from the subject. After examining the velocity contours and vector arrays, the cough jet center was defined as the midpoint of the cough, where the greatest velocities are present (a typical cough is shown in Figure 3 ). If no jet center was evident, the cough was not where V s is the residual velocity within the chamber at t = 0, and V peak is the maximum velocity at the peak of the cough. where t peak is the time at which V peak is observed. HWA was used to measure the instantaneous velocity magnitude at single location. A constant temperature anemometry (CTA) unit was attached to a probe containing a tungsten wire 1.25 mm long and 5 µm in diameter, and voltage readings were recorded at a rate of 1 kHz. The system was calibrated for low airflow speeds using a specialized facility detailed in Mohamed. 38 The sensor was located x = 1.00 m downstream from the mouth of the participant, y = −0.17 m below the inlet centerline, since preliminary trials had indicated that the cough jet traveled along a slightly downward trajectory. A moving average filter was applied to the measurements to filter out high-frequency fluctuations and noise. Two separate CFD models have been previously developed, based on large eddy simulation (LES) and unsteady Reynolds-averaged Navier-Stokes (URANS) methods, 36 in which the boundary conditions were selected based on the average volumetric flow rate profiles and mouth opening diameters determined experimentally by Gupta et al. 21 The inlet velocity was specified to match that velocity (shown in Figure 5 ), and a mouth diameter of D = 0.0217 m was used for the simulation. The LES data were filtered using the same moving-averaged method utilized to smooth the HWA data. The recruitment procedures were approved by Western's Research Ethics Board (REB approval no. 108945). Subjects with influenza-like illnesses were referred to the study after making an appointment to see a physician at Western Student Health Services. Following referral, subjects completed a questionnaire to determine whether they were eligible to participate in experiments. Participants needed to be between 18 and 35 years of age, and in the last 24 hours, they should have experienced a fever and cough/sore throat in the absence of Throughout the duration of the investigation, 77 sets of experiments were conducted for 58 different subjects. Table 1 summarizes the number of measurements taken for each method. In the 2014 investigation, the HWA probe was only used as a reference, and it was not calibrated. PIV data were not available during the 2017 study period. The was measured, which is consistent with the spread angles previously determined. 21, 22 Through a critical review and analysis of experimental and numerical investigations, it was determined that steady free jets also exhibited spreading angles between 20° and 35°, which was consistent with the findings of the present investigation. 41 The peak velocity magnitude, V peak , was used to characterize the strength of cough flow behavior in the far field. A comparison of the strengths of coughs from different cohorts is displayed in Table 3 . As expected, the greatest average velocities were noticed at the cough center, followed by the HWA location. The centerline location had the fewest number of coughs with usable data. Very often, the cough missed this location entirely, or the peak velocity at the location was below 0.20 m/s. A t test with a significance level of α = .05 did not show a statistically significant difference in the peak cough velocity magnitude, V peak , between each cohort. Similar values were noticed for the average peak instantaneous velocities at the cough center, at HWA location, and along the inlet centerline (Tables 4 and 5 To better understand the relationship between the peak jet center velocity and the time at which it occurs, the two quantities are plotted in Figure 8 , and a reciprocal fitting curve was obtained where k is a mathematical fitting parameter related to the average distance between the front of the jet and the point at which the maximum velocity occurs (k = 0.55 m for the cough jet center (PIV)). Although there is scatter in the data, a user may choose a V peak value, obtain τ from Figure The fluctuations about the moving average of the HWA signal were used to characterize the turbulence present at the peak of the cough. The normalized power spectral density (PSD) was used to indicate the distribution of turbulent energy. Figure 9 displays the normalized power spectral density for coughs for an example cough. A slope of −5/3 was observed between 8 and 100 Hz, which is consistent with the Kolmogorov decay law for the inertial subrange, and only subtle differences are observed between coughs. The turbulence scales in this region are larger than those of viscous dissipation, but smaller than those in the energy-containing region. The turbulence intensity was estimated from Equation 6 (I ua ) and from integrating the area under the PSD curve (I us ). where V ′ rms is the root mean square of velocity fluctuations. The turbulence intensity, computed from the residual fluctuations about the moving average and averaged across all coughs, was I ua = 8.9%. As expected, this was very similar to the turbulence intensity computed from the power spectrum, and a value of I us = 8.4% was calculated (4.6% difference). The validity of the numerical simulation was determined by comparing the normalized experimental velocity profiles with those determined computationally ( Figure 10) . A peak velocity of 0.84 m/s is obtained from the URANS simulation, whereas it is higher in the LES with a value of 0.90 m/s. As expected, the peak velocity obtained from PIV experiments was slightly higher (1.17 m/s), since it was calculated from the instantaneous peak velocity magnitude, whereas the CFD values come from window-averaged time histories. Both the URANS and LES showed good agreement with the experimental results. The LES may have been a better representation of a realistic cough, but it was substantially more computationally expensive and so may not be practical for some investigations. To gain a better understanding of the decay in V peak with respect to distance (x) from the source, the instantaneous velocity time histories were extracted along the simulated cough centerline at 0.1-m intervals, starting at the inlet and extending to x = 1.5 m. The normalized peak velocity, V peak-n , was calculated according to Equation 7, where V peak is the peak velocity magnitude observed at each location, and V inlet is the peak velocity magnitude observed at the inlet, and it is shown vs distance from the inlet in Figure 11 . and mouth diameter at the origin from Gupta et al, 21 the cough jet velocity will decay to 1% of its initial velocity at x = 115D. Since droplets and droplet nuclei smaller than 10 µm can remain suspended for long periods of time, 36 it is likely that they will be transported by the jet and there is potential for the transmission of viruses at these distances. In practice, the ventilation currents within a room will play a more dominant role at such low velocities. In this study, the near field was considered to be the region close to the mouth, where the flow field was primarily determined by the initial impulse of the cough, while the far field was defined as the region beyond this where the flow field resembled that of a transient turbulent free jet. In the far field, the velocity time history profiles were similar when normalized using the local peak velocity and timescale. Figure The biological air sampling method used in this study was unable to assess the relative significance of far-field virus transmission. This could mean that that three forced coughing events do not produce a significant quantity of viral droplets, or it is possible that the virus was present but not sampled due to the small area of the sampling cassettes. There is also a possibility that the air sampling method used destroyed the viral particles upon impaction, and, perhaps, in future investigations, aerosols should be sampled for more coughs or an alternative sampling method should be used. It is possible that the type of viral infection will influence the production of mucous within the lungs and may alter the way a person will cough, such that it may not be accurate to consider all cough jets containing respiratory viruses as being identical. It is also assumed that the trajectory of a natural cough may be different from those studied here since the motion of the head is restricted in the trials. Another assumption used for this investigation is that a forced cough will behave aerodynamically like a naturally occurring cough. By combining experimental measurements using particle image velocimetry and hot-wire anemometry and data obtained from Although a high degree of variability is observed between individual coughs, the average profiles are a good representation of the expiratory airflow field produced by a human cough in the region x ≥ 15D. This approximate location can be used to distinguish between near-field region, where the flow field is primarily governed by the initial cough impulse, and the far-field region, where the locally No statistically significant differences were observed in the velocity or turbulence characteristics between coughs from sick or healthy participants, and evidence was provided to support the claim that velocity data obtained from healthy participants in F I G U R E 11 Variation of normalized peak velocity and distance from the inlet (CFD) previous studies can be used to approximate the flow field of coughs from individuals who have been infected with respiratory viruses. We would like to express our deepest gratitude toward the members of the Advanced Fluid Mechanics Research Group at the University of Western Ontario, especially Dwaipayan Sarkar, for his assistance with conducting the experiments. We would also like to thank our colleagues at Sunnybrook Research Institute for performing the viral analysis. Finally, we would like to thank the staff at UWO Student Health Services, for their assistance with the recruitment of participants. A systematic review of the social and economic burden of influenza in low-and middle-income countries Global role and burden of influenza in pediatric respiratory hospitalizations, 1982-2012: a systematic analysis Severe acute respiratory syndrome-related coronavirus-The species and its viruses, a statement of the Coronavirus Study Group World Health Organization. Coronavirus disease 2019 (COVID-19). Situation report-23 Droplet fate in indoor environments, or can we prevent the spread of infection? Indoor Air How far droplets can move in indoor environments -revisiting the wells evaporation-falling curve Aerosol transmission of influenza A virus: a review of new studies Airborne spread of expiratory droplet nuclei between the occupants of indoor environments: a review The size and concentration of droplets generated by coughing in human subjects Exhaled droplets due to talking and coughing Size distribution and sites of origins of droplets expelled from the human respiratory tract during expiratory activities Modality of human expired aerosol size distributions Cough aerosol in healthy participants: fundamental knowledge to optimize droplet-spread infectious respiratory disease management Exhalation of respiratory viruses by breathing, coughing, and talking Measurements of airborne influenza virus in aerosol particles from human coughs Viable influenza A virus in airborne particles expelled during coughs versus exhalations. Influenza Other Respir Viruses Relationship between expired lung volume, peak flow rate and peak velocity time during a voluntary cough manoeuvre Evaluation of cough using digital particle image velocimetry Study on transport characteristics of saliva droplets produced by coughing in a calm indoor environment Characterization of expiration air jets and droplet size distributions immediately at the mouth opening Flow dynamics and characterization of a cough A Schlieren optical study of the human cough with and without wearing masks for aerosol infection control Particle image velocimetry of human cough A new methodology for studying dynamics of aerosol particles in sneeze and cough using a digital high-vision, high-speed video system and vector analyses Violent expiratory events: on coughing and sneezing Western cold and flu (WeCoF) aerosol study preliminary results Detection of airborne severe acute respiratory syndrome (SARS) coronavirus and environmental contamination in SARS outbreak units Measurement of airborne influenza virus in a hospital emergency department Concentrations and size distributions of airborne influenza A viruses measured indoors at a health centre, a day-care centre and on aeroplanes Influenza virus transmission is dependent on relative humidity and temperature High temperature (30°C) blocks aerosol but not contact transmission of influenza virus Transmission of influenza virus in a mammalian host is increased by PB2 amino acids 627K or 627E/701N Transmission of influenza virus via aerosols and fomites in the guinea pig model Preventing the transmission of avian or pandemic Influenza in health care facilities with limited resources: Learning resource package. Jhpiego: An affiliate of John Hopkins University Infection Control for Health Care Providers. Albany: Thomson Delmar Learning A numerical investigation of human cough jet development and droplet dispersion Infectious virus in exhaled breath of symptomatic seasonal influenza cases from a college community Experimental measurements of far-field cough airflows produced by healthy and Influenza infected human subjects Thesis Experimentation, Validation, and Uncertainty Analysis for Engineers Methods for sampling of airborne viruses The flow field in turbulent round free jets Experimental investigation of far-field human cough airflows from healthy and influenzainfected subjects