Airborne Transmission and Control of Influenza and Other Respiratory Pathogens

Abstract

Despite uncertainty about the specific transmission risk posed by airborne, spray-borne, and contact modes for influenza, SARS-CoV-2, and other respiratory viruses, there is evidence that airborne transmission via inhalation is important and often predominates. An early study of influenza transmission via airborne challenge quantified infectious doses as low as one influenza virion leading to illness characterized by cough and sore throat. Other studies that challenged via intranasal mucosal exposure observed high doses required for similarly symptomatic respiratory illnesses. Analysis of the Evaluating Modes of Influenza Transmission (EMIT) influenza human-challenge transmission trial—of 52 H3N2 inoculated viral donors and 75 sero-susceptible exposed individuals—quantifies airborne transmission and provides context and insight into methodology related to airborne transmission. Advances in aerosol sampling and epidemiologic studies examining the role of masking, and engineering-based air hygiene strategies provide a foundation for understanding risk and directions for new work.

1. Introduction

Seasonal and pandemic influenza remain global threats. Seasonal flu kills up to 650,000 people each year and pandemics have the potential to cause millions of deaths and disrupt societies. Despite surpassing the 100-year anniversary of the 1918–1919 global influenza pandemic with a death toll estimated at over 50 million, present-day non-pharmaceutical prevention strategies—including engineering controls like germicidal ultraviolet technology (GUV), filtration, and ventilation—remain inadequately used to quell seasonal influenza epidemics and emerging pandemics as demonstrated with ongoing epidemiologic waves of COVID-19. Stringent social isolation remains an effective approach over the centuries but may only achieve population-level compliance for short periods of time. Testing, vaccination, and therapies are helpful but have not been available at the outset of emerging pandemics, and face issues of waning effectiveness as pathogens evolve, and logistical and social issues related to rapid production and equitable dissemination. It is widely appreciated that the quest for improved non-pharmaceutical controls and vaccines is dependent upon knowledge of influenza virus transmission via direct contact, large droplet spray, and aerosol inhalation and deposition along the respiratory tract. Increasing precision and confidence of quantified risks posed by airborne and other transmission modes support better design and evaluation of engineering controls and other strategies to reduce population spread. To rapidly identify airborne pathogens and continually update knowledge about airborne infection potential of evolving pathogens, there is a need for sentinel epidemiologic and bioaerosol sampling surveillance systems.

Influenza intervention trials showed that the use of hand hygiene and surgical masks to reduce contact and large droplet exposure resulted in only mild risk reduction among susceptible household contacts of influenza cases and may have facilitated more airborne transmission [1]. Human challenge studies have shown that infection initiated through aerosols, compared with nasal instillation [2, 3], required a lower dose and resulted in more severe disease. Inhalation of bioaerosols is likely important for other acute, viral, and respiratory infections and was convincingly implicated by airborne viral transport computational fluid dynamic models for a deadly SARS-coronavirus outbreak [4, 5, 6]. The capacity to directly measure the extent and intensity of transmission risk posed by bioaerosols represents uncertainty for which research is needed. Failure to quantify the contribution of exhaled bioaerosols impedes the advocacy for and effective use of control measures and facilitates population vulnerability during seasonal epidemics and pandemics.

William Wells described the quantum theory of airborne infection [7] whereby infection risk is described by exposure to infectious doses, or quanta (which is, more specifically, the dose that would infect 63% of those exposed), generated by infectious individuals over time. Studies quantifying influenza virus and SARS-CoV-2 virus shed into exhaled breath aerosols using a Gesundheit-II (G-II) bioaerosol sampler support an understanding of airborne contamination by infectious individuals, provides a way forward for precisely estimating airborne infection risk in terms of virions with infectious potential and genome copies measured by quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) [8, 9, 10, 11]. Human challenge transmission trials offer one way forward to quantifying the airborne transmission risk between infectious and susceptible individuals with known levels of inhalation exposure to exhaled breath, paired with measurements of viral load in exhaled breath. Real-world study in congregate settings where exposure may be unavoidable, including in healthcare or other public gathering places offers another approach that may provide more generalizable findings, yet may be more logistically challenging to achieve valid estimates of exposure [12]. Yet emerging genomic sequencing methods that can identify viral mutations shared between epidemiologically linked individuals can confirm transmission chains and may offer clues to the specific mode of transmission [13].

Reliable prediction of airborne risk informs disease control efforts by providing information about the relationship between various levels of exposure control via engineering controls—air disinfection by GUV, ventilation, and filtration—and reduce airborne transmission. This information is needed to inform public health approaches and infrastructure design to provide appropriate air hygiene for mitigating emerging pandemic viruses before effective vaccines and therapies become available. This chapter provides an overview of airborne viral infection dynamics and control with a focus on the scientific underpinnings required for future epidemiologic study designs under which longitudinal surveillance of contact networks can revolutionize understanding of airborne infection transmission by pinpointing transmission routes and refining estimates of infection risk by airborne and other modes in indoor spaces. Results from this line of research provide key information for guiding the strategic use of prevention methods—especially air disinfection by GUV—to protect against seasonal epidemics and pandemics in shared air spaces and, in particular, among immunologically vulnerable populations.

2. Disease burden and public health impact

The Forum of International Respiratory Societies emphasizes that acute respiratory infections are the greatest contributors to the global disease burden, responsible for 4 million deaths annually. CDC reports that influenza resulted in 9–36 million illnesses and up to 56,000 deaths each year since 2010 in the US, with annual estimated direct and indirect costs of $87 billion [14]. Respiratory infections cost over $15 billion annually in the UK [15]. Globally, seasonal influenza kills up to 675,000 people each year and influenza pandemics have the potential to cause millions of deaths and severe societal disruption. The health and economic burdens are amplified in developing nations with less access to health services [16]. The devastating loss of life, the morbidity and economic losses from COVID-19, trends in spillover of pathogenic avian influenza to humans with pandemic potential, and an increasingly interconnected world, all create an urgent case for improved prevention methods.

Prevention of these substantial population health threats cannot rely solely on vaccines, which are often poorly matched to rapidly evolving strains. During pandemic, the lag time in the production and dissemination of vaccines leads to widespread vulnerability and underscores the need for interventions based on viral exposure reduction to interrupt transmission. It is widely appreciated that the quest for improved non-pharmaceutical prevention methods—including reducing exposures through building-level air disinfection—or social distancing, and vaccines are dependent on understanding transmission risk via contact, large droplet spray, and fine-particle aerosol respiration (i.e., airborne) [17]. Vaccine development benefits from an understanding of host-pathogen dynamics related to transmission mode. There is strong evidence supporting the critical role of airborne transmission, and it is well-recognized that infection initiated by the airborne route is likely to cause more severe symptoms compared to infections initiated by contact or large droplet spray [2, 3].

The US CDC typically has recommended protective behaviors such as washing hands, covering coughs, and donning masks to reduce contact and droplet exposure, but has provided little specific guidance related to air cleaning or respiratory protection (e.g., fit tested N95s) to mitigate aerosols that are capable of penetrating and circumventing surgical or cloth masks. Intervention trials showed that the use of hand hygiene and surgical masks to hinder contact and droplet exposure resulted in only mild risk reduction among susceptible household contacts of nearly 800 influenza cases and may have promoted a greater proportion of airborne transmission [1]. Furthermore, those most likely to be exposed to airborne influenza, due to the use of hand hygiene plus surgical mask, tended to present with more severe symptoms characterized by fever and cough. Despite some delays in intervention initiation and imperfect adherence, such trial conditions reflect realistic population usage, while randomization and robust sensitivity analyses support internal validity to the extent possible. Although implementation of masking and engineering controls such as GUV, filtration, and ventilation are well supported by existing evidence, the state of the science benefits from investigation to better quantify airborne transmission risk and the extent of the effectiveness of GUV, filtration, and ventilation.

A clear, dose-response relationship between dormitory rebreathed air fraction and likelihood of retrospective, self-reported acute respiratory infection (ARI) was observed in a study of 3,712 students in Tianjin, China [18]. A separate airborne infection risk model suggested that increased clean air supply could effectively control population spread of ARIs including influenza but may not have much effect on highly contagious infections like measles [19]. However, this study used estimated values of influenza contagiousness based on an airplane outbreak [20], where there was uncertainty about outdoor air exchange, and all secondary influenza cases were assumed to be connected to the index. More recently a comparison of two university dormitories in Maryland, USA showed that compared with a dormitory with higher ventilation the dormitory with low ventilation had 4 times (95% confidence 0.69–163.02) the ARI rate, although the sample size of infections reported in the high ventilated dormitory reduced the ability to make more conclusive comparisons [21].

While modulating airflow and ventilation can influence airborne contamination quantities and human exposure, unequivocal evidence from exposure chambers demonstrates the inactivation of aerosolized respiratory pathogens including influenza [22], vaccinia virus [23], and TB [24] under exposure to upper-room 254 nm UV-C light (GUV), representing a highly effective strategy to interrupt airborne transmission. But whereas current control techniques are unlikely to be strategically deployed, improved characterization of risk by transmission mode enables the most effective use of existing control strategies and may provide health benefits knowledge to help catalyze investment by communities, government, and public health agencies.

3. The human-challenge transmission trial for quantifying infection modes

A meeting of globally recognized influenza transmission experts was convened by CDC in 2010 to address knowledge gaps about the relative importance of influenza transmission modes that are reflected in uncertainty about hospital care and general population prevention guidelines [25]. The meeting discussed possible animal and human transmission experiments and explored the possibilities of conducting epidemiological studies with engineering and/or personal protective interventions. Although there was great enthusiasm for studies of population infection surveillance with upper room GUV or other airborne control interventions, preliminary work in this area was lacking. Ultimately it was determined that a human challenge-transmission study with interventions to control for transmission mode, surveillance of aerosol shedding, environmental conditions, comparison of aerosol infectivity of experimental and naturally infected influenza cases would represent the most scientifically sound approach.

4. Aerobiologic pathway for influenza and other respiratory infections

An abundance of laboratory evidence substantiates the aerobiologic pathway for influenza and other ARIs and supports new epidemiologic studies of transmission. The aerobiologic pathway [26], consists of a) generation of particles containing infectious microbes from the respiratory tract or environmental sources, b) maintenance of infectivity and persistence in the air before reaching a susceptible host, and c) deposition in at least one vulnerable locus in the respiratory tract of the new host.

With respect to infectious particle generation, exhaled breath particles contain a respiratory fluid lining of the small airways and are generated by small airway closure and reopening [27, 28, 29]. A team led by Milton at the University of Maryland observed 218 half-hour exhaled breath samples from 142 symptomatic influenza cases and detected culturable influenza virus in 39% of fine-particle aerosols (≤5 μm) with geometric means of 37 infectious particles by fluorescent focus assay and 3.8x10⁴ RNA copies by qRT-PCR (geometric standard deviations 4.4 and 13, respectively) [11]. Using a G-II bioaerosol collection device to sample natural breathing (including incidental coughs), this research clearly shows that influenza cases can generate many virus-laden particles. The same research team using a similar methodology detected SARS-CoV-2 in 36% of fine and 26% of coarse aerosols, while also detecting infectious viruses [10]. Others using the G-II showed that singing produced the highest proportion of positive fine aerosols, followed by talking and breathing [30].

Once generated, infectious aerosols maintain infectivity and persist in the air before reaching a susceptible host. The airborne movement of infectious particles has been implicated in human and animal transmission of influenza and other respiratory pathogens. Computational fluid dynamics and multi-zone models simulating a three-dimensional aerosol plume rising upwards and around an apartment building with a SARS-coronavirus index case predicted the location of secondary cases [4]. Noti and colleagues measured infectious influenza in aerosols that had traveled across a room [31]. Upward dispersion of aerosols with slow settling velocity has been confirmed by influenza. A transmission between infected guinea pigs housed >100cm below exposed animals [32]. Numerous ferret studies report similar results. The ability for airborne particles to travel and initiate disease was implied by two postal workers who became infected with Anthrax following a known release of spores and no other known exposures [33].

Biologically active airborne particles carry public health significance given the potential for prolonged suspension and scenarios of exposure before removal occurs or through recirculated air that has not been filtered or sterilized. Studies of biological decay in aerosolized virus maintained in a rotating drum demonstrated infectious potential for influenza [34] and coronavirus [35] after 23 hours and 6 days, respectively. Although the exact sizes of the laboratory-generated aerosols used were not reported, these studies demonstrate prolonged infectiousness in particles <10 μm. The rate of biological decay as a function of temperature and relative humidity has been characterized through laboratory manipulation of viral-laden droplets [36]; and through airborne simulations with bacteriophage Phi6, a surrogate for influenza and coronaviruses [37]. Reduced decay corresponded with lower droplet salt concentrations associated with high and low vapor pressures, consistent with epidemiologic observation of peak transmission during the hot and rainy season in the tropics, and the cold and dry season in temperate climates. However other research using aerosolized virus from human airway epithelial fluid suggests that the influenza virus remains infectious independent of relative humility [38]. This latter work may be more convincing given the use of a more realistic human model. The aerosol half-life of SARS-CoV-2 has been reported at 1.1 hours (95% CI 0.64–2.64) [39], with infectivity measured at 16 hours with potential for longer persistence under longer observation [40].

Inhalation of airborne virus and deposition at a vulnerable locus in the respiratory tract can initiate infection. A human challenge study demonstrated an infectious dose for inhaled influenza A aerosols as low as 0.6–3 TCID₅₀ [2]. A study of exhaled breath from confirmed influenza cases showed that 99 and 87% of particles were less than 5 and 1 μm, respectively [28]. This shows that exhaled breath aerosols are well within the size range to penetrate the lower lung. Fine particle aerosols exhaled from naturally infected influenza cases have been shown to carry infectious viruses [9, 11]. Given that, epidemiologic, laboratory, and challenge studies fail to definitively confirm human airborne transmission and produce valid risk models, there is a need for methods that maximize external validity to community settings and enable confirmation of transmission modes for a range of ARIs. Observation of community transmission provides an ideal platform to validate risk models that parameterize the aforementioned aerobiologic path—viral aerosol generation, persistence, and deposition—leading to valid estimation of infectious dose. Observation of exposed, asymptomatic individuals satisfies the concerns of Fraser and colleagues, which identified asymptomatic cases as key to pushing R₀ above one [41].

5. Studies of influenza transmission risk by mode and the anisotropic hypothesis

Hand hygiene and face masks have been assessed for their potential to reduce influenza transmission and gain information about transmission mode-related risk. Cluster-randomized trials with hand hygiene and facemask interventions found mild reductions in risk among intervention users (effect for hand hygiene and facemask groups, separately) that did not reach statistical significance [42]. This finding was consistent with those from studies performed in Hong Kong and Bangkok that showed the effect of hand hygiene plus facemask to be small at best [1, 43, 44]. A similar result was observed for crowded, urban households in upper Manhattan after 19 months of follow-up in 509 households [45]. However, a meta-analysis showed that hand hygiene plus facemask interventions were associated with a statistically significant 27% reduction in transmission risk [46]. Hand hygiene alone had no significant effect but showed a trend toward reducing risk under higher humidity and suggesting a predominance of aerosol transmission in temperate climates that is weakened in tropical climates. Given that facemasks have been assessed to reduce viral RNA copies contained in coarse aerosols by 25-fold and fine aerosols by 2.8-fold [9], if such reductions are associated with reduced transmission risk, then the meta-analysis findings make sense. Several other studies and review papers provide extensive evidence for the role of airborne particles in both influenza [5, 47, 48], and SARS-CoV-2 [49, 50, 51, 52, 53, 54, 55] transmission.

The hypothesis that influenza is anisotropic—that the route of transmission influences disease presentation [5]—is supported by early studies of human exposure to influenza contained in aerosols and nasal droplets [2, 3, 56], where aerosol exposure was more likely to result in influenza-like disease characterized by fever and cough, compared with nasal mucosa exposure representative of contact and droplet routes. The community-infected cases documented by Knight and colleagues exhibited similar symptomatology as Alford’s infected volunteers, suggesting a natural tendency toward aerosol transmission. These findings were more recently borne out in ferrets where aerosol-infected animals not only presented with more severe symptoms but also shed more virus than their nasally-inoculated counterparts [57]. Similarly, cynomolgus macaques exposed to SARS-CoV-2 via aerosols were more likely to experience fever and severe respiratory pathology compared with those exposed via intratracheal/intranasal drops, suggesting similar anisotropy [58].

6. Findings from EMIT human challenge transmission trial

The human challenge-transmission trial (Evaluating Modes of Influenza Transmission [EMIT], ClinicalTrials.gov number NCT01710111) was designed to achieve an expected 40% SAR, however, achieved an actual SAR of 1.3% [59]. This finding on its own fails to provide definitive results regarding transmission modes, yet the low transmission rate from close-quarters exposure of infectious influenza cases over four consecutive 12–16-hour days with sero-susceptible individuals suggests that the contact and spray-borne transmission modes were not important contributors. Comparison of this result with the proof-of-concept study that achieved a SAR of 8.3% under much lower exposure time and ventilation motivates discussion about the role of ventilation and exposure to airborne pathogens [60].

Bueno de Mesquita and colleagues used CO₂ data from the transmission trial, and knowledge of aerosol viral shedding by experimentally infected primary cases (known as “viral Donors”) and applied the rebreathed-air equation—a modification of the Wells-Riley equation—to estimate an infectious quanta generation rate and RNA copy number per infectious quantum [19]. This analysis showed that the particular group of exposed individuals where the single secondary infection was observed was among the group with the highest exposure to virus contained in the exhaled breath of the Donors to which they were exposed. This suggests that the transmission may have occurred through the airborne mode. Assuming this, the airborne quanta generation rate (q) (95% CI) for influenza in the controlled human transmission trial environment among infected Donors and airborne viral shedding Donors was estimated to be 0.029 (95% CI 0.0270, 0.03) and 0.11 (0.088, 0.12) per hour, respectively. The number of RNA copies per infectious quantum was 1.4E+5 (95% CI 9.9E+4, 1.8E+4). Given this quantum generation rate, and levels of viral shedding in a college campus community in dormitory rooms evaluated for exhaled breath exposure, the typical viral shedder presents a low risk of transmission to a susceptible roommate during three nights of exposure in a well-ventilated dormitory but a moderate risk in a poorly ventilated dormitory. Supershedders at the 90th percentile of fine aerosol shedding would present high risk even in the higher ventilated dorm. The effect of higher ventilation could be modeled using the rebreathed-air equation and typically points towards the need for levels of air exchange far beyond what might be achievable by ventilation alone, underscoring the importance of air disinfection by GUV and filtration to mitigate superspreading.

The next question is whether the EMIT human volunteers experimentally infected by intranasal droplets simulate naturally-acquired infections to a comparable degree. To address this question, the EMIT study included an investigation of community influenza cases presenting with influenza-like illness. There was a low probability artificial nasal inoculation would have resulted in the highest levels of symptom severity and viral shedding observed among naturally infected cases selected on the basis of febrile illness [61]. Findings from these analyses generate new knowledge about influenza infection, disease, and transmission and inform future studies aimed at improving our understanding of respiratory infection transmission dynamics and associated disease. There is limited data elsewhere about the extent of shedding as a function of symptom profile, although asymptomatic individuals have been shown to shed 1–2 log₁₀ RNA copies fewer than symptomatic influenza cases [62]. The extent to which asymptomatic infections may be more representative of populations infected by upper respiratory mucosal exposure is unclear.

The computed infectious quantum generation rate enables the comparison between estimated exposure to influenza virus and infection risk. Thus, given levels of exhaled breath aerosol viral shedding and ventilation rates for indoor shared air spaces, the Wells-Riley equation can be applied to estimate infection risk. Of course, this assumes that the assumptions inherent in the computation of the q in the EMIT human challenge-transmission trial can be generalized to other transmission scenarios. The population of susceptible volunteers had low HAI and MN titres, representing above-average susceptibility to the general population, suggesting q may be overestimated. The computed q must also be interpreted with caution because it represents a point estimate, with confidence bounds generated by empirical bootstrap, given that it was derived from a single transmission event. The q for influenza in the challenge trial is relatively low compared with the few estimations done for other respiratory infections. Yet applying the EMIT-derived RNA copy to infectious quantum relationship to naturally infected influenza cases shedding the most virus among 142 mostly healthy young adults gave a q value of 630 [63]. Analysis of a super spreading event on an airplane suggested q or 100 for influenza virus, while q for rhinovirus has been estimated at 4 [19]. Careful epidemiologic investigation of an explosive measles outbreak in an elementary school showed q of over 5,000 [64]. Analysis of a SARS-CoV-2 outbreak in a poorly ventilated restaurant yielded an estimated q of about 80 [55], while a super spreading event at a choir practice led to an estimated q of over 900 [65].

7. EMIT trial limitations and questions for future work

The EMIT challenge-transmission trial, like Alford’s challenge study with aerosol viral exposure, used a population with low pre-existing antibodies to the challenge virus subtype. Thus, these studies are useful for demonstrating transmission dynamics with susceptible secondary cases but lack generalizability to the general population with varying levels of immunity. That only one transmission event was observed in a Control Recipient represents a major limitation, as the mode of transmission cannot be well deduced, and the risk ratio represents the lower bound for infection risk and lends uncertainty to the confidence bounds. Nonetheless, the EMIT analyses attempted to a) learn what was possible about influenza transmission given that the study was unique in its design and the largest human-transmission trial conducted to date, and b) fully assess the limitations of the study design to inform future investigations. Numerous questions exist to drive future studies aiming to refine risk assessment and optimize population prevention strategies. Such questions include:

1. To what extent do climate, subtype, and engineering controls affect transmission risk while accounting for variability in human susceptibility to illness, and the concurrent deployment of administrative and behavioral interventions?

2. What is the optimal set of social contact networks and measured indoor air quality variables to predict observed transmission events in a contained community setting (e.g., dormitories, military barracks, nursing homes, hospitals, schools, occupational settings)?

3. What epidemiologic strategies would best support the effective identification of exposure networks and confirm transmission events, potentially by infection mode?

4. To what extent does infection mode influence the immune response and subsequent viral shedding peak, temporal shedding patterns, and virion infectiousness, and to what extent does age, sex, prior infection, vaccination, and immune status modify these effects?

5. Could viral load quantities and community characterization sampled at different sites (lung-produced aerosols versus upper respiratory tract-produced mucosa) and times throughout the course of infection indicate the mode by which infection was initiated?

6. How do symptoms correlate with infectious viral shedding from the lung versus upper respiratory mucosa across age, sex, socio-behavioral factors, immune status, and subtype?

7. To what extent is aerosol shedding a function of viral concentration in the respiratory fluid of the distal airways and expiratory volume, and how might these relationships change with cough, speech, singing, and other drivers of expiratory particle generation?

8. Implications for study design

Findings from the experimental challenge-transmission model should be evaluated in studies of real-world epidemiology and population transmission dynamics. This way the potentially important contributions of other important variables can be assessed: a) immunity and shedding dynamics, b) socio-behavioral factors related to human-human interaction and exposure, c) overall well-being including psychological stress, sleep, physical activity, and diet [66], d) features of the built environment where exposure occurs including temperature, humidity, sanitary ventilation (combination of outdoor airflow, filtration, and GUV), and e) the role of other airborne exposures including particulate matter, ozone, and nitrogen oxides.

The advantage of the experimental trial in a controlled environment is that a relationship can be drawn between viral shedding quantity and subsequent secondary attack rates, giving a dose-response relationship. However recent advances in genomic sequencing and bioinformatics show a path forward for using molecular markers, in combination with epidemiological contact and exposure surveillance, to confirm transmission chains [67, 68, 69]. Sequences from identified transmission pairs may be able to give information about infection mode if viral communities evolve distinctly in the lung versus the upper respiratory mucosa. There is evidence that influenza may manifest as compartmentalized infections in the lung and nasopharynx [11, 70]. Airborne transmission likely involves viral communities produced in the lung, while contact transmission likely involves nasal communities, thus enabling a path to identify infection route that requires characterization. Considering the nasopharynx and lung as separate entities that carry the ability to infect independently, reconstruction of transmission chains in observed contact networks may be possible by analyzing shared variants [13]. Bayesian approaches can be used to infer transmission events for outbreaks that are not completely sampled and/or are ongoing [71].

A study that simultaneously monitored ventilation rates in two neighboring dormitories housing first and second-year students and respiratory infections among the population found a trend towards a higher infection rate in the dormitory with lower ventilation, suggestive of a relationship between inhalation exposure and infection risk [21]. This study also showed a gradient of exposure levels to exhaled breath between rooms in a corridor that could support epidemiologic investigation of transmission chains. Longer studies with larger populations should be done that combine contact investigation and sequence analysis to confirm transmission chains. Symptom assessment, specimen collection for quantification of mucosal and exhaled breath viral load, viral community, immune biomarkers, and other health-related factors related to stress would provide necessary data sources to assess the relationship between exposure and infection risk that could be modified by immunologic factors.

9. Conclusions and implications for public health practice

Although new studies are needed to refine estimates of transmission risk by various modes to understand the relationships between infection mode, dose, symptoms, age, sex, immunity, and environment, the existing state of knowledge is sufficient to support the scientific underpinnings of public health interventions aimed at reducing transmission and population epidemics through targeted airborne exposure control. At the very least, airborne infection preventive measures should be used as part of precautionary strategies to protect populations from loss of life and livelihood associated with emerging pandemics. In the case of influenza, it may be that the infectious generation rate of the average infectious aerosol shedder is low enough to pose an only mild risk under conditions with abundant sanitary ventilation but may pose moderate to severe risk under conditions of less sanitary ventilation (Figure 1). Fine particle aerosol supershedders may pose a substantial risk regardless of sanitary ventilation. Although they may be quite rare in the population, supershedders may account for most of the population spread as shown in the case of SARS-CoV-2 [72, 73]. Investigation of exactly how much of this risk can be attenuated by engineering controls opens the door for well-informed exploration of building design and operation strategies. That sanitary ventilation measures provide contribute to the control of any airborne transmitted pathogen represents a major advantage.

Figure 1.

Factors contributing to elevated risk of airborne infection transmission. Supershedders emitting high rates of infectious aerosol through exhaled breath or another aerosolization mechanism (e.g., aerosolization from toilet flush) can lead to high risk alone, or average spreaders can generate high risk under conditions of low sanitary ventilation or among populations of immunologically susceptible populations.

The magnitude of infection control measures required to prevent the community spread of airborne contagion is related to the infectivity of the pathogen (i.e., infectious dose shedding rate). Testing, quarantine, and isolation are critical measures to interrupt transmission by removing exposures to infectious sources and should always be considered. Yet there are challenges with achieving widespread access to sensitive tests for emerging pathogens, and compliance with quarantine and isolation procedures. Engineering controls including GUV, filtration, and ventilation provide an effective layer of protection that can be facilitated by the government as a social good requiring little if any behavior change or compliance at the population scale. Engineering controls—with an emphasis on GUV when dealing with highly infectious pathogens—can help move societies beyond reliance on social isolation and masking, especially given the social fatigue with these measures observed after more than two years of COVID-19 pandemic [74]. Vaccines and therapies are important measures, despite some problems with social acceptance. Yet, they take time to develop and deploy widely and may wane in effectiveness as pathogens evolve, and thus their use holds little bearing on the importance of engineering controls for population protection. Yet all available control measures can help as part of a layered approach and may be required for extremely infectious agents.

As demonstrated by Nardell and colleagues in the case of TB [75], and Bueno de Mesquita and colleagues in the case of influenza [63], there exist potential limits to the extent that ventilation controls alone can control transmission risk in shared air environments. Given that seasonal influenza epidemics cause substantial burden of morbidity and mortality, and COVID-19 and other emerging pandemic pathogens can exact an even more devastating toll, there is a great opportunity for wider use of infection control via engineering controls with the greatest effectiveness. SARS-CoV-2 subvariants appear to be increasing in infectivity and a super spreading event suggests that a highly infectious case could produce over 1,000 quanta per hour [65]. This compares with measles cases which may shed 500 to over 5,000 quanta per hour [19, 64] and an influenza supershedder who generated approximately 600 quanta per hour [63]. Yet upper-room GUV has been shown to mitigate measles spread in elementary schools [76, 77, 78] and provide many times the equivalent sanitary ventilation provided by outdoor airflow and filtration [23, 79, 80]. Newer far-UVC applications that allow safe exposure to the light directly have been shown to have similar levels or better air disinfection with potential for widespread use in a greater variety of settings [81, 82]. The use of GUV technology offers to reduce the theoretical limits of sanitary ventilation and lowers transmission risk in congregate settings where filtration, outdoor airflow, and masking may not offer sufficient protection.