Article Text

Download PDFPDF

Coughs and sneezes spread diseases: but do ‘aerosol generating’ procedures?
  1. David T Arnold1,
  2. Fergus W Hamilton2,3,
  3. Ed Moran4
  1. 1 Academic Respiratory Unit, University of Bristol, Bristol, UK
  2. 2 MRC-IEU, University of Bristol, Bristol, UK
  3. 3 Microbiology, North Bristol NHS Trust, Westbury on Trym, UK
  4. 4 Department of Infectious Disease, North Bristol NHS Trust, Bristol, UK
  1. Correspondence to Dr Ed Moran, Department of Infectious Disease, North Bristol NHS Trust, Bristol BS10 5NB, UK; Ed.moran{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The COVID-19 pandemic has led to the widespread use of continuous positive airway pressure (CPAP) and high flow nasal oxygen systems (HFNO) for respiratory support, with the former recently shown to reduce the requirement for intubation and mechanical ventilation, easing pressure on intensive care and improving outcomes.1

One major concern remains the theoretical risk for aerosolisation of SARS-CoV-2, with both CPAP and HFNO currently considered ‘aerosol generating procedures’ (AGPs). This classification, endorsed by Public Health England, the WHO and US CDC, dichotomises medical procedures into those that potentially generate infectious aerosol (smaller particles, generally <5 µm with the potential to remain suspended for extended periods) and those that do not.2 Enhanced infection control precautions (eg, airborne mask protection, increased ventilation, cohorting) are thus advised when patients undergo AGPs but are not mandated outside these settings.

The evidence supporting this concept of AGPs is largely based on epidemiological observation, particularly from the SARS pandemic, with limited physical data supporting the role of aerosol generation in most AGPs.3 Studies conducted over the course of the pandemic are undermining the AGP paradigm, with convincing evidence that the aerosol generated in normal respiratory activities exceeds the limited amount generated by (nearly) all tested AGPs.4 5 To date, few studies have attempted to measure actual infectious virus during suspected AGPs, given the difficulties of performing these measurements.

In this month’s journal, Winslow and colleagues provide useful data on the aerosolisation of SARS-CoV-2 by air sampling while patients are undergoing both HFNO and CPAP.6 Thirty patients admitted with COVID-19 were included in this observational study. By co-enrolling participants who were taking part in RECOVERY-RS,1 a trial randomising individuals to different respiratory support modalities, they reduced confounders between the groups. They compared three groups of 10 participants, receiving either supplemental oxygen via Venturi facemask, CPAP with standard viral filter, or HFNO with a typical flow rate of 50–60 L/min. Using an identical sampling protocol for all participants, they performed RT-PCR assays for human and SARS-CoV-2 RNA from the air and immediate clinical environment. They found that, while 21/30 (70%) participants tested positive for SARS-CoV-2 RNA in their nasopharynx, only 30% had a positive result from the sampled air with associated lower viral loads (higher Ct values). From the surrounding surfaces, more viral RNA was identified with 46% having at least one positive sample, with the floor being the most common surface affected. Crucially, there was no difference in viral RNA detected between the participants in terms of the type of respiratory support being used. It is of note that, although viral culture was attempted on all PCR positive samples, no air samples and just one nasopharyngeal sample were culture positive. Patients were sampled an average of 12 days after symptom onset, so this is perhaps unsurprising, given the high Ct values and dynamics of SARS-CoV-2 infection in immunocompetent patients.7

This study adds to the mounting evidence that AGP classification is unhelpful in defining risks of transmission. It has resulted in overcautious measures for certain settings, mandating full PPE for all intubations4 and preventing relatives visiting the sickest patients, while underplaying the risk in others such as emergency department staff swabbing newly symptomatic patients, ward staff providing personal care for patients with known COVID-19 and non-clinical staff cleaning contaminated floors.

The results here also correlate with mechanistic studies of aerosol generation from respiratory support procedures that have shown that breathing, speaking and coughing are more aerosol generating than CPAP or HFNO.5 8 This, and the decreasing infectivity of COVID-19 patients over the course of an admission, goes some way to explaining the low rates of SARS-CoV-2 seropositivity among intensive care staff compared with other patient-facing staff groups,9 10 despite early evidence suggesting higher viral loads in such patients.

In summary, it is clear that aerosol transmission of SARS-CoV-2 does occur and there is mounting evidence that the risk of infection in hospital settings does not relate well to historic AGP guidance. A certain level of risk has always been accepted within healthcare settings. Throughout the pandemic, national policies have sought pragmatically to minimise rather than eliminate all risk, adapting throughout the pandemic following the initial classification of SARS-CoV-2 as a high consequence infectious disease. However, the evidence now suggests that the balance of risk is placed in exactly the wrong place, with enhanced PPE mandated in those situations in which aerosol is least likely and people least infectious—for example, those working in ITU or caring for those receiving CPAP—rather than in those where aerosol is most likely and people most infectious—such as coughing patients with early infection in admitting units and general wards.

Insisting on airborne precautions for all healthcare staff in contact with possible SARS-CoV-2 patients will be hugely costly—and, as any user of such PPE knows, uncomfortable—for increasingly limited gains if transmission of SARS-CoV-2 falls and potential risks of infection are attenuated by vaccination. Contrariwise, removing all airborne precautions is unlikely to be considered acceptable by those working in areas used to such precautions, and would seem imprudent given emerging evidence of waning immunity and the prospect of increasingly transmissible variants. Perhaps only one thing is clear: there are nuanced decisions ahead for those tasked with deciding infection control policy.

Ethics statements

Patient consent for publication



  • Twitter @gushamilton

  • Contributors DTA and FWH produced a first draft. EM then extended and developed it.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed.

Linked Articles