Cough in Simulated Clinical Settings

Simulation studies have tended to utilize either human volunteers or artificial simulators as the source of cough during experiments. The most applicable study identified from the literature search was performed in a chamber (to model a medical examination room) with a cough simulator used to generate an aerosol-laden cough and aerosol particle counters located at different positions within the room. ^{Reference Lindsley, King and Thewlis20} With the study specifically focused on the aspect of infectious bioaerosols dispersed by patients in a health care environment and the risk to health care workers, aerosol particles with diameters of 0.3µm to 7.5µm were evaluated with results showing that cough-generated aerosol particles became rapidly dispersed throughout the room after just five minutes. As with any cough-simulator, a limitation of using machinery as the cough source is the inability to replicate the impact of buoyancy. ^{Reference Lindsley, King and Thewlis20} The naturally heated human cough plume is usually warmer than the ambient air, hence its buoyancy, and although this may not have a significant impact on larger particles, it is likely to have a significant effect on smaller particle sizes. ^{Reference Lindsley, Reynolds, Szalajda, Noti and Beezhold21} Additionally, cough simulators are unable to replicate the same real-world mechanisms of aerosol generation – primarily being shear stress as airflow meets the mucous membrane, vibration between structures in close proximity, and bronchial fluid burst on terminal airway reopening. ^{Reference Dhand, Jie and Li22}

A methodologically similar study to Lindsley, et al ^{Reference Lindsley, King and Thewlis20} investigating the spread of the influenza pathogen during coughing concluded that aerosol transmission likely plays a role in the spread of influenza. ^{Reference Noti, Lindsley and Blachere23} Whether the results of studies looking at a specific pathogen can be generically applied to the airborne transmission of other infections is debatable, but Noti, et al’s ^{Reference Noti, Lindsley and Blachere23} study attempts to quantify infectivity from a coughing event which is relevant to other ARIs. Studies attempting to provide clarity regarding infectivity report the presence of airborne ribonucleic acid when reviewing both influenza and coronaviruses, but they rarely found viable viruses in the air. ^{Reference Shiu, Leung and Cowling24} The SARS-CoV-2 virus has been detected in the air with a half-life of just over one hour, ^{Reference Van Doremalen, Bushmaker and Morris1} and this has numerous citations within the evidence base as proof of “viable” virus. However, Van Doremalen, et al’s ^{Reference Van Doremalen, Bushmaker and Morris1} study was laboratory based with an aerosolized environment created in a Goldberg drum, so it has a significant limitation of not being representative of real-world data. Following initial ambiguity, the UK government now recognizes that SARS-CoV-2 is transmitted via the airborne route, ²⁵ and the research broadly supports this concept. ^{Reference Wilson, Norton, Young and Collins12}

Surgical Mask as a Source Control Device during Cough

Face mask efficacy is the primary focus of the human volunteer studies. Air flow leakage is an aspect that is often also reported during these studies where face masks are used as a source control device (ie, being worn by the source of the cough). Of the studies using human volunteers, a significant proportion used the Schlieren technique – a well-established method to visualize the flows of gases and liquids by use of differences in light refraction. ^{Reference Settles26} However, this technique does not provide data on aerosol size, concentration, or mass distribution, so it is limited in inferring risk relating to aerosol transmission. The evidence shows considerable lateral air leakage around a surgical mask, ^{Reference Tang, Liebner, Craven and Settles27–Reference Lindsley, Blachere, Law, Beezhold and Noti30} with differing assertions on whether this effect redirects the jet-stream to a less harmful direction ^{Reference Tang, Liebner, Craven and Settles27} or should be considered as a major hazard for those in the vicinity of the cough. ^{Reference Viola, Peterson and Pisetta28} Researchers agree that a surgical mask is effective as a source control device but there is discrepancy in the degree of effect with a surgical mask recorded as blocking anywhere between 59% ^{Reference Lindsley, Blachere, Law, Beezhold and Noti30} and ∼90% ^{Reference Hamilton, Gregson and Arnold2,Reference Cappa, Asadi, Barreda, Wexler, Bouvier and Ristenpart31} of aerosols produced by coughing. It is noteworthy that the study reporting poorer efficiency used a manikin head and cough simulator to test mask performance. With the UK guidance relating to the COVID-19 pandemic encouraging patients to wear a face mask during transportation in an ambulance, ⁶ the element of face mask leakage is pertinent when weighing up risk to health care workers in a confined clinical setting such as an ambulance. Significantly, the evidence shows that loose-fitting face masks do not effectively prevent aerosol emissions contaminating the surrounding environment. ^{Reference Leung, Chu and Shiu32}

Cough Variation Amongst Individuals

Mathematical modelling studies are increasingly viewed as useful tools in clinical research, with the tendency being to use modelling when systematic reviews fail to adequately answer research questions. The results of modelling studies can be considered indicative with findings often determined by the validity of the primary data applied. The key parameter applied to modelling studies, and that which differs amongst the evidence, is the exhaled microdroplet/aerosol particle distribution and estimated viral copies produced during a cough. Variation amongst individuals has resulted in a “low” and “high” emitter range often being adopted with one modelling study finding that coughing emissions ranged between 0.000277 copies/cm³ (low emitter) to 36,030 copies/cm³ (high emitter) with the PM10 (particle size below 10µm) accounting for approximately one-half of these values. ^{Reference Riediker and Tsai33} The aerosol number produced by a single cough has a range of between 900 to 300,000 particles when measuring aerosols between 0.35-10µm. ^{Reference Lindsley, Pearce and Hudnall34} The SARS-CoV-2 virus is thought to be 60-140nm in size ^{Reference Zhu, Zhang and Wang35} and the virion particles have an affinity to attach to larger particles in the 0.3-10µm range, ^{Reference Ganann, Kitila, Patel, Brook and Pisegna36,Reference Leung and Sun37} so it is entirely plausible that the virus could be contained within particles of the size range associated with aerosols. Riediker and Tsai’s ^{Reference Riediker and Tsai33} study used research based on healthy individuals ^{Reference Yang, Lee, Chen, Wu and Yu38} to form their modelling of microdroplet dispersal and application of previous research in this way highlights a significant flaw that can be directed towards modelling studies as infected individuals are thought to produce a higher concentration of aerosol. ^{Reference Hamilton, Gregson and Arnold2,Reference Lindsley, Pearce and Hudnall34}

Aerosol Emissions from Coughing whilst Infected

A case-control study found that particles per cough in infected (influenza) versus non-infected participants were reported as 75,400 and 52,200, respectively. ^{Reference Lindsley, Pearce and Hudnall34} The study reported particle size distribution for one participant considered to be a high emitter, displaying a generic increase in all size ranges when infected. This type of analysis would be particularly useful for the SARS-CoV-2 pathogen due to the inhibitory impact the virus has on surfactant production caused by the virion binding to ACE2 receptor sites and subsequently infecting and damaging Type II alveolar cells. ^{Reference Piva, Diblasi and Slee39} Surfactant acts to reduce alveolar surface tension and is known to increase aqueous elasticity. ^{Reference Johnson and Morawska40} In application to the SARS-CoV-2 infection, bronchiole fluid film burst may occur more frequently and at smaller diameters generating a larger volume of aerosols in the lower particle size range.

Although noting some significant limitations relating to reporting bias, Hamilton, et al ^{Reference Hamilton, Arnold and Bzdek14} cites a similar theme regarding aerosol particle number concentration from coughing when infected with SARS-CoV-2 as to that highlighted for influenza by Lindsley, et al. ^{Reference Lindsley, Pearce and Hudnall34} Hamilton, et al ^{Reference Hamilton, Arnold and Bzdek14} recruited hospitalized COVID-19 patients as a case cohort (n = 8) alongside a control cohort (n = 25) of healthy volunteers. Using optical and aerodynamic particle sizers, the volunteers underwent protocolized procedures, including coughing. The environment in which the research was carried out also represented a significant limitation, differing between the groups due to logistical constraints with an ultra-clean laminar flow operating theatre used for healthy volunteers and a negative pressure ventilated room used for hospitalized volunteers. Nevertheless, the study concluded that aerosol number concentration was higher during cough for the infected volunteers.