In early January 2020, news filtered through to the general public of a disease outbreak caused by a novel coronavirus centred around a live animal market in Wuhan, China. A media release from New Zealand’s Ministry of Health on 24 January 2020 noted the virus caused pneumonia.[[1]] It advised the public to “take steps to reduce their risk of infection”, including by “regularly washing your hands, covering your mouth and nose when you sneeze”, staying home when sick and “avoiding close contact with anyone with cold or flu-like symptoms”. These risk reduction measures assumed the virus spread via close contact, contaminated surfaces and large droplets of saliva emitted during coughing and sneezing. These assumptions aligned with longstanding teaching within the international Infection Prevention and Control (IPC) community that respiratory viruses generally spread via large respiratory droplets that fall rapidly to the ground within 1–2 metres of the source (“droplet transmission”).  

By March 2020, aerosol scientists were publicly arguing that SARS-CoV-2 and other respiratory viruses also spread via tiny respiratory droplets that remain suspended in the air for longer periods (“airborne transmission”).[[2]] Significantly, they noted such tiny droplets (“aerosols”) are emitted during normal breathing and talking, even without coughing, sneezing or “aerosol generating procedures”. This understanding subsequently helped explain several observations about the pandemic, including indoor super-spreading events; instances of long-range transmission; and the tendency of the virus to transmit during the pre-symptomatic phase of infection.

Unfortunately, the World Health Organization (WHO) was initially reticent to acknowledge the expertise of non-clinical aerosol scientists and explicitly recognise SARS-CoV-2 as an airborne pathogen, delaying important IPC mitigation measures in both healthcare and community settings.[[2]]

Science of airborne disease transmission

Any respiratory activity (including shallow breathing) emits particles of various sizes defined loosely by droplet size and aerodynamic properties. At two ends of the spectrum, are small droplets (aerosols) that float and large droplets that rapidly fall to the ground under gravity.

The content of the droplets depends on their origin within the respiratory tract. They consist principally of saliva, hydrated mucus and/or lung surfactant, meaning they are mostly water with some carbohydrates, proteins and salts. They may carry virions, in proportion to the concentration of virions in the fluids from which they originate. When exhaled, the droplets change in size in relation to temperature and humidity, and tend to shrink by evaporation, leaving small low-water-content particles (“droplet nuclei”).[[3]]

When infected with a respiratory pathogen, a person may generate exhaled droplets in the lung and conducting airways, or in the upper airway (trachea, mouth, pharynx and nasal passages). Breathing, speaking, shouting, singing, coughing and sneezing can generate more droplets of larger average sizes. Large droplets tend to fall faster than they evaporate and cluster around the source. Small droplets are also more concentrated near the source but can be coughed or sneezed several metres and drift in the air for up to several hours.[[3]]

Historically, the IPC literature has distinguished between particles which are droplets (diameter >5 microns) or aerosols (diameter <5 microns). Aerosol scientists have always seen this threshold as inaccurate and unhelpful. Wells’ original research in this area generated an “evaporation falling curve” and placed the division at 100 microns.[[4]]

This is the largest particle size that in appropriate environments can remain suspended in the air for more than five seconds and be inhaled.[[3]] Generally, respiratory droplets follow the exhaled air, but also settle towards the ground under gravity. Settling may be slowed, or prevented altogether, by up-draughts of air. Wells observed large droplets (>100 microns) tended to settle faster and fall under gravity within two metres of the source. This led to the recommendation for two-metre distancing between people when infection transmission via large respiratory droplets is a concern.[[5]]

In the last three years, researchers and clinicians have increasingly recognised that most SARS-CoV-2 transmission occurs via aerosol transmission (<100 microns). This occurs when an infected person exhales virion-containing aerosols, which mix with the ambient air, and a susceptible person close to or distant from the infectious case inhales them. The probability of exposure to an infectious dose depends on many factors including the viral load of the source; the rate of aerosol production; proximity and duration of exposure; the recipient’s immune status; inhalation dose (which masking at source and recipient may mitigate); and the rate of dilution with clean air through ventilation. Infectious dose dilution is very quick in the outdoor environment. Indoors, dilution may occur through mechanical (fan-based) ventilation systems, open windows, humidity, temperature, and air movement and mixing.[[3]]

A pictorial summary is shown in Figure 1. Based on this new understanding of airborne transmission of SARS-CoV-2, recommended IPC measures have extended to include precautions to limit airborne transmission. No single IPC measure will provide 100% protection from infection. A key principle is to mitigate infection risk with multiple layers of protection, such as immunisation and public health measures which reduce the frequency or duration of contact with infectious people—this includes physical distancing, use and type of mask, ventilation levels and air-cleaning technologies.

View Figures 1–2 and Tables 1–2.

Modelling airborne transmission risk

Modelling can estimate the risk of a susceptible person developing an infection from inhaling droplets an infectious person has exhaled. It involves estimating the total mass and active virus concentration of virus-carrying particles exhaled (related to viral load); dilution by mixing with clean air; settling out of droplets on surface; inactivation of virus by time, ultraviolet (UV) light or other means; the rate at which the susceptible person inhales air; and the effect of masking by either person. The probability of infection developing must also be calculated using a dose-response model.

Several groups have developed risk estimation tools based on these methods. One of the most detailed is the “Airborne Infection Risk Calculator”.[[6]] The most common is an exponential dose-response model, defining an “infectious quantum”, which is the number of viable ribonucleic acid (RNA) copies required to start an infection in 63% of susceptible people. As the infectious quantum and the susceptible person’s vulnerability are usually uncertain, they represent the greatest uncertainty in this method. The method is more robust when used to calculate odds ratios between two different scenarios, such as well-ventilated vs poorly ventilated rooms.

Where exhaled breath can be assumed to mix immediately and uniformly throughout the room, the Wells–Riley formulation can be used to calculate the quantity of virus inhaled.[[7]] This assumption is reasonable over long periods in rooms with much air motion. In practice the risk of infection is higher when people are in close proximity, inhaling each other’s breath before it is diluted with ambient air. Better estimates come from using computational fluid dynamics (CFD) to model air flow and mixing, which can resolve jets of breath and air currents caused by ventilation and heat sources. Uncertainties remain as the air motion at any given time is highly variable, but CFD can yield insight into the importance of close-range vs long-range infection—see Figure 2 for an example of its use in the Managed Isolation Quarantine Facilities (MIQF) ventilation assessments.

New Zealand’s contribution to the growing understanding of airborne transmission

New Zealand had border restrictions in place from March 2020. Everyone entering the country had to quarantine for 14 days in hotels that were functionally converted into MIQF. These facilities had extensive processes, protocols and rules for physical distancing between guests. During their stay, guests routinely underwent polymerase chain reaction (PCR) testing of nasopharyngeal swabs at set intervals and if symptoms developed. All positive swabs underwent whole-genome sequencing (WGS). This arrangement effectively allowed for a natural observational experiment as WGS identified all transmission events, which then enabled targeted and highly thorough investigations into how and when transmission occurred. Tools for investigation included routinely collected records of guest and staff activity; CCTV footage; interviews; and key-card data (giving the precise time whenever a guest re-entered their room). Using all of this information, the investigators generated hypotheses on when and how transmission most likely occurred.

In October 2020, two transmission events from a guest cohort to Christchurch MIQF nurses were identified. Work records and interviews narrowed down possible transmission events to brief interactions at each source case’s doorway. In each event the nurse was following protocol: wearing full personalised protective equipment including a standard ear-loop medical mask, eye protection, gown and gloves. In one case, two senior IPC nurses observed the interaction and identified no breaches in process. Throughout the interaction, which lasted 40–60 seconds, the asymptomatic source case stood in the doorway wearing a medical mask and remained silent. The nurse removed their wristband and replaced it with another of a different colour. This was the only contact the nurse had with any potential source case with a matching WGS profile. The investigators concluded that airborne transmission facilitated by poor ventilation was the most likely mechanism, bypassing the standard medical masks the two nurses wore. They thought the nurses were probably exposed to a sudden wave of air heavily contaminated with infectious aerosols from each source’s room soon after they opened their door.[[9]]

Given these findings, investigators reviewed an earlier transmission event in September 2020, when one guest infected another in the adjacent room. The first investigation had attributed the infection to fomite transmission through a virally contaminated lid of a shared rubbish bin in the corridor outside both guests’ rooms. However, re-examining the evidence made it clear that airborne transmission relating to opening of adjacent doors in rapid succession was far more likely.[[10]]

Other transmission events in MIQF around the country were investigated—some with the use of modelling described above to test plausibility—and in most cases airborne transmission was found to be the most likely explanation. Interdisciplinary collaboration was critical to understanding these transmission events.[[11]]

Merging clinical and aerosol scientist expertise

On 28 March 2020, WHO stated that except for “aerosol generating procedures”, SARS-CoV-2 was not airborne. In July 2020, 239 aerosol scientists published an open letter calling on the medical community to recognise airborne spread of SARS-CoV-2. The authors noted that recognition of airborne transmission had substantial implications for preventative public health measures including improving indoor ventilation; air cleaning (by filtering or disinfection); avoidance of indoor crowding; and masking.[[12]]

In March 2021, a WHO-funded systematic review of the evidence for airborne transmission stated, “the lack of recoverable viral culture samples of SARS-CoV-2 prevents firm conclusions to be drawn about airborne transmission”.[[13]] The key pitfall of this review was that the evidence underpinning the existing paradigm of “droplet transmission” was not given the same level of critical scrutiny or even examined. A month later, “Ten scientific reasons in support of airborne transmission of SARS-CoV-2” was published.[[14]] The authors urged clinicians and policy makers to act, rather than waiting for somewhat arbitrary laboratory-level proof that would be difficult to obtain. Their preferred precautionary approach would assume airborne transmission has occurred until proven otherwise.

As more observational case studies, mathematical modelling and experimental studies supporting airborne transmission accumulated, the WHO’s communications began to implicitly support this message. Yet it was not until December 2021 that its website explicitly stated both short- and long-range transmission of SARS-CoV-2 was important.[[15]]

Environmental controls for airborne diseases

Protective measures against airborne transmission can involve source control (reducing viral dispersion from the index case) or transmission control (reducing the likelihood of non-infected people inhaling the virus) (see Table 1).

Practical responses to new information and understanding

The observations of airborne spread in New Zealand’s MIQFs led to a revision of IPC practices, starting in Canterbury in late 2020 and rolling out quickly to other centres. Staff masks were progressively upgraded to N95s, ventilation engineers were employed to assess every MIQF, and air cleaners with HEPA filtration were strategically located in “dead air” spots such as elevators and corridors. Routine surveillance testing frequency increased to identify and move asymptomatic infectious cases to appropriate isolation earlier. These rapid changes in response to the new paradigm were enabled by strong leadership from the clinical staff involved in MIQF.[[16]]

How to protect the community from transmission of SARS-CoV-2 now

In late 2022 it is well understood that SARS-CoV-2 is predominantly spread by airborne transmission. Masking, particularly of the infected person (source control), is very effective at reducing transmission. The more people who wear masks, the greater the impact. For greatest impact, everyone should be masked in crowded and/or poorly ventilated indoor public spaces, although this is not always achievable or reasonable.[[17]] Additional measures to prevent transmission are needed.

Consider the analogy of potable drinking water. Just as the majority of New Zealanders can access clean water from a tap without having to personally filter and disinfect it, so too should people be able to trust that the air they breathe is clean. Many public buildings achieve around 1–2 air changes an hour (ACH) when the aim should be a minimum of 4–6 ACH.[[18]] Encouragingly, researchers in Hong Kong have shown that improved ventilation of a room can significantly reduce long-range and short-range transmission of respiratory pathogens.[[19]] Table 2 presents these and other measures for improving indoor air quality, which from a health perspective is both achievable and desirable.

Dilution with fresh air is favoured where sufficiently high flow rates can be achieved and comfortable temperatures and noise levels maintained. If windows cannot be opened, ducted ventilation systems can often be adjusted or upgraded to achieve greater dilution, although building occupants may have limited control over the system a landlord installs. Carbon dioxide (CO{{2}}) monitors are inexpensive and give an immediate assessment of the fresh air supply rate relative to the number of occupants.[[20,21]] This monitoring has additional benefits given CO{{2}} itself is a hazard in high concentrations, affecting cognition.[[22]] Many CO{{2}} monitors also measure temperature and humidity, which help building occupants learn to balance fresh air and heating or cooling to maintain a comfortable, healthy environment. Our experience is that using CO{{2}} monitors for even one week can develop new healthy ventilation habits.

Where sufficient fresh air flow is unattainable, air cleaning can be effective in reducing virus and particulate matter, although it does not reduce CO{{2}} levels as fresh air does. It can be achieved by filtration (HEPA or similar filters) installed in either ducted ventilation systems, which can require expensive installation work, or portable units. The latter are relatively inexpensive and quick to install although when bought in bulk capital costs can mount up and maintenance costs (power and filter cleaning and/or replacement) need to be budgeted for. The major limitations associated with portable air filters are ensuring physical placement is appropriate, reliance on occupants to turn them on and, sometimes, noise levels can be problematic.[[23]] Guidance to assist in choosing appropriate air cleaners is available, for example https://cleanairstars.com/.  Those filters which introduce reactive species into the air to break down pathogens (e.g., ionizers and hydrogen peroxide systems) are not recommended without careful risk assessment, as the reactive species may potentially be harmful if inhaled.[[24]] If sufficient ACH cannot be achieved with air filters another alternative is upper air or in-duct UV germicidal irradiation (UVC, UGVI) although unlike filtration it does not remove particles such as soot, which might in themselves be hazardous.[[25]]

Poor air quality leads to health and wellbeing issues beyond infectious diseases. A large body of research documents concerns with volatile organic compounds, build-up of CO{{2}} and many other air pollutants. In one such report, published (ironically) in January 2020, the Royal College of Physicians in the United Kingdom recommended that the government should establish a cross-government committee to coordinate working to improve indoor air quality in public sector buildings and residential homes.[[26]] More recently, similar recommendations have been made in New Zealand and the United Kingdom.[[27,28,29]]

Guidance is available on how to assess air quality in New Zealand classrooms using CO{{2}} meters along with methods to improve the air quality.[[30]] This guidance is generic and transferable to many public and private buildings.

In the short-term similar accessible information guides with advice that is easy to implement needs to be provided to the business sector including hospitality and entertainment. Longer term, an equity-based approach would target investment and education in environments which house poorly vaccinated populations, e.g., pre-schools and primary schools. Monitoring of CO{{2}} levels in public buildings where people congregate in groups such as entertainment venues, cafeterias, education facilities, whare kai and churches should be normalised. Appropriate responses to high CO{{2}} levels and education about using ventilation (ducted and natural) and air cleaning to achieve comfortable, safe indoor environments should be part of the commissioning process in all buildings used by the public.

In the longer term, the Government needs to lead with building codes in residential and commercial sectors that treat clean air provision with as much importance as earthquake safety.

The removal of non-pharmaceutical public health interventions (such as mask wearing requirements in public spaces) makes the use of engineering controls to minimise exposure to contaminated air even more important.[[31]] Improving indoor air quality will not only reduce COVID-related illness, but all morbidity related to poor air quality. With this in mind, it’s understandable that reliably maintaining high indoor air quality standards has been described as the new “sanitation”.[[32]]