key: cord-0289890-tibgo381
authors: Thornton, G. M.; Fleck, B. A.; Fleck, N.; Kroeker, E.; Dandnayak, D.; Zhong, L.; Hartling, L.
title: The impact of heating, ventilation, and air conditioning design features on the transmission of viruses, including the 2019 novel coronavirus: a systematic review of ultraviolet radiation
date: 2021-10-14
journal: nan
DOI: 10.1101/2021.10.12.21264904
sha: 599e85eb96f78fa6f33da7e00b61551f5ccb9b34
doc_id: 289890
cord_uid: tibgo381

Respiratory viruses are capable of transmitting via an aerosol route. Emerging evidence suggests that SARS-CoV-2 which causes COVID-19 can be spread through airborne transmission, particularly in indoor environments with poor ventilation. Heating, ventilation, and air conditioning (HVAC) systems can play a role in mitigating airborne virus transmission. We conducted a systematic review of the scientific literature examining the effectiveness of HVAC design features in reducing virus transmission; here we report results for ultraviolet (UV) radiation. Following international standards for systematic reviews, we conducted a comprehensive search and synthesized findings from 32 relevant studies published between 1936 and 2020. Research demonstrates that: viruses and bacteriophages are inactivated by UV radiation; increasing UV dose is associated with decreasing survival fraction of viruses and bacteriophages; increasing relative humidity is associated with decreasing susceptibility to UV radiation; UV dose and corresponding survival fraction are affected by airflow pattern, air changes per hour, and UV device location; and UV radiation is associated with decreased transmission in both animal and human studies. This comprehensive synthesis of the scientific evidence examining the impact of UV radiation on virus transmission can be used to guide implementation of systems to mitigate airborne spread and identify priorities for future research.

COVID-19, the disease caused by a coronavirus (SARS-CoV-2), was declared a pandemic by the World Health Organization in March 2020. 1 Since then, public health authorities worldwide have sought evidence about the route of transmission and appropriate public health measures to mitigate virus spread. Certain viruses have been proven capable of transmitting via an aerosol route. 2 In the case of aerosol transmission, virus-laden aerosols are expelled by humans and remain airborne for extended periods of time. Emerging evidence suggests that the SARS-CoV-2 virus can spread through airborne transmission under certain circumstances, particularly in indoor environments with poor ventilation. 3, 4 Selecting appropriate measures to protect the occupants of indoor spaces based on informed, interdisciplinary research is critical to managing the spread of infectious disease. 5 Heating, ventilation, and air conditioning (HVAC) systems can play a role in mitigating the airborne transmission of viruses by removing or diluting contaminated air inside a building enclosure where humans breathe. [5] [6] [7] Many features within HVAC systems can influence transmission, such as ventilation rates, filters, humidity, and ultraviolet (UV) radiation. Under ultraviolet germicidal irradiation (UVGI), a dose of UV light is delivered to the aerosolized virus which causes damage to the DNA impeding its ability to replicate. The ability for these cells to infect a host are therefore lost. 8 UV radiation can be applied within mechanically ventilated spaces in the building environment through in-duct or upper-room lamp fixtures. Irradiation in an enclosed space, such as in-duct UVGI, allows for better control of the UV dose, resulting in better control of particle/pathogen exposure to UV radiation. Non-enclosed systems, such as upper-room UVGI, depend on air circulation to drive the particles/pathogens to an irradiated zone near the ceiling (designed to shield virus transmission. The insight drawn from this review could help answer questions of the utility of UVGI as an adjunct technology to curb the spread of the SARS-CoV-2 in mechanically ventilated indoor environments. Further, understanding effectiveness relative to technology set-up and UV dose could inform control measures.

This paper describes the results of a systematic review to identify and synthesize the scientific We developed an a priori protocol 21 that is publicly available and the systematic review 22 is registered. We followed standards for the conduct of systematic reviews defined by the international Cochrane organization 23 with modifications for questions related to etiology. 24 We report the review according to accepted reporting standards. 25

A research librarian (GMT) searched three electronic databases (Ovid MEDLINE, Compendex, Web of Science Core) from inception to June 2020 using concepts related to virus, transmission, and HVAC. The search strategy for Ovid MEDLINE appears in Table 1 ; the strategies were peerreviewed by two librarians (TL, AH) prior to implementing the searches. The search was updated in January 2021. We screened reference lists of all relevant papers as well as relevant review articles. We identified conference abstracts through Compendex and Web of Science; abstracts . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ;  https://doi.org/10.1101/2021. 10.12.21264904 doi: medRxiv preprint were not included but we searched the literature to see whether any potentially relevant abstracts had been published as complete papers. We did not limit the search by year or language of publication; however, we only included English-language studies due to the volume of available literature and resource constraints. References were managed in EndNote with duplicate records removed prior to screening.

Study selection occurred in two stages. First, two reviewers independently screened the titles and abstracts of all references identified by the electronic databases searches. Relevance of each record was classified as Yes, No or Maybe. Conflicts between Yes/Maybe and No were resolved by one reviewer. We conducted pilot testing with three sets of studies (n=199 each) to ensure consistency among the review team. After each set of pilot screening, the review team met to discuss discrepancies and develop decision rules. The second stage involved two reviewers independently reviewing the full text articles and applying the inclusion/exclusion criteria. Studies were classified as Include or Exclude. Conflicts between Include and Exclude were resolved by consensus of the review team. Conflicts between different exclusion reasons were resolved by one reviewer. We pilot tested the second stage of screening with three sets of studies (n=30 each). After each pilot round, the review team met to resolve discrepancies. We conducted screening using Covidence software. Table 2 lists our inclusion and exclusion criteria. As noted above, this systematic review was part of a larger effort to examine different HVAC design features and virus transmission. We searched and screened for all design features at once, but only studies evaluating UV radiation are synthesized here. While our interest was UV within HVAC systems, we also included studies of . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ;  https://doi.org/10.1101/2021. 10.12.21264904 doi: medRxiv preprint upper room UVGI because of its similar utility and mechanism of air disinfection. We searched for a variety of agents but prioritized studies of viruses or agents that simulated viruses; we planned to include other agents (e.g., bacteria, fungi) only if studies were not available that were specific to viruses. We included studies of bacteriophages, which are viruses that infect bacterial cells. 19 We were interested in studies of the indoor built environment (e.g., office, public, residential buildings) which had mechanical ventilation. We included primary research that provided quantitative results of the correlation or association between installed UV radiation and virus survival or transmission. We placed no restrictions on year of publication; we included only English-language, peer-reviewed publications.

For experimental studies, we assessed risk of bias based on three key domains: selection bias, information bias and confounding. [26] [27] We assessed each domain as high, unclear, or low risk of bias using signaling questions 28 from guidance documents for the different study types we included; e.g., animal studies, laboratory experiments, epidemiological studies. [26] [27] 29 For modelling studies, we assessed the following three key domains: definition, assumption, and validation. 29-30 Definition considers model complexity and data sources, assumption considers the description and explanation of model assumptions, and validation considered model validation and sensitivity analysis. 30 Also, we assessed each domain as high, unclear or low risk of bias based on signaling questions. 29-31 The risk of bias items were pilot tested among three review authors, then two reviewers (GMT, BAF) applied the criteria independently to each relevant study and met to resolve discrepancies.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021. 10.12.21264904 doi: medRxiv preprint We extracted general information about the study (authors, year of publication, country of corresponding author, study design) and methods (setting, population [as applicable], agent studied, intervention set-up). We extracted details on UV treatment parameters (where available), including: wavelength; UV dose; exposure time; and fluence rate. Also, we extracted information (where available) regarding relative humidity (RH). The studies were grouped as "in-duct UVGI"

and "upper-room UVGI." We extracted quantitative data, as well as results of any tests of statistical significance related to UV features. A priori, our primary outcome of interest was quantitative measures of the association between UV radiation and virus transmission; however, during the review we realized that most studies focused on proxy variables such as virus survival. Therefore, we extracted data on actual transmission where available (i.e., infections), as well as proxy variables (e.g., survival fraction (SF), dose-response of UV dose and survival fraction, susceptibility (Z), and equivalent air changes per hour (ACH) due to UV radiation (ACHuv)). represented as SF = exp(-ZD), where Z is the susceptibility and exp() represents exponential function. Equivalent ACH due to UV radiation (ACHuv) is the number of air changes per hour (ACH) that would produce the same reduction in virus concentration as obtained using UV radiation. We created a data extraction form spreadsheet to ensure comprehensive and consistent capture of data. One reviewer extracted data and a second reviewer verified data for accuracy and completeness. Discrepancies were discussed by the review team.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021. 10.12.21264904 doi: medRxiv preprint We anticipated that meta-analysis would not be possible due to heterogeneity across studies in terms of study design, UV features examined, outcomes assessed, and reporting of results. We developed evidence tables describing the studies and their results (as reported by the authors of the primary studies). We provide a narrative synthesis of the results of relevant studies. To allow for meaningful synthesis and comparison across studies, we divided the studies into four groups: aerosolized virus, modelling, animal studies, human studies. Within the aerosolized virus group, the effect of RH was further examined.

The electronic searches and other sources yielded 12,177 unique citations; 2,428 were identified as potentially relevant based on title/abstract screening and 568 met the review's inclusion criteria ( Figure 1 ). Of the 568, 125 were relevant to UV radiation and, of those 125, 32 were relevant to UV radiation and virus ( Figure 1 ). Among the 32 relevant studies there were: 16 aerosolized virus and bacteriophage studies (Table 3) , 7 modelling studies (Table 4 ), 4 animal studies (Table 5) , and 5 human studies (Table 6 ). Studies were published between 1936 and 2020 (median year 2007.5).

While the majority of the experimental and modelling studies were published between 2005 and 2020, with one exception in 1964, the human studies are all from the 1940s and the animal studies spanned from 1936 to 2020. The majority of studies were conducted in the United States (n=24).

Studies were funded by national research funding organizations (n=13), industry (n=6), a university and state grant (n=1), and hospital (n=1); 2 studies reported no external funding and 8 studies did not report funding source.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint Table 3 shows that 17 viruses and five bacteriophages from 16 studies were inactivated by UV radiation. Generally, susceptibility was determined from the dose-response relationship of UV dose and survival fraction 11, 13, 15, [32] [33] [34] ; however, Walker and Ko 10 calculated susceptibility from a single dose and corresponding survival fraction, and Lin et al 35 used both approaches. Some entries in Table 3 Increasing UV dose was associated with decreasing survival fraction for 10 viruses and five bacteriophages from 12 studies. Increasing UV dose was associated with decreasing survival fraction where the dose-response relationship was used to calculate susceptibility for seven studies (Table 3) . 11, 13, 15, [32] [33] [34] [35] Additionally, UV dose was associated with decreasing survival fraction where dose varied by exposure time, 36,40 number of UV fixtures, 12 and fluence rate. [41] [42] Increasing RH was associated with decreasing susceptibility in four studies 11,13,32-33 for a variety of infectious agents including viruses (Influenza A, Vaccinia virus, PRRSV) and bacteriophages (MS2, phiX174, phi6, T7) ( Figure 2 ). Cutler et al 33 reported that PRRSV susceptibility was significantly lower at ≥80%RH compared with 25%RH-79%RH. In addition, four studies that report susceptibility at one RH are included in Figure 2 10, 15, [34] [35] where three viruses were coronaviruses (murine hepatitis virus (MHV) coronavirus, human coronavirus 229E, human coronavirus OC43). Considering the findings for influenza A 13,34 and coronaviruses, 10,15 UV radiation inactivated these enveloped, single-stranded RNA viruses and increasing UV dose was . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint associated with decreasing survival fraction characterized by the susceptibility. If enveloped, single-stranded RNA animal viruses behave like influenza A (Figure 2 ), then increasing RH may be associated with decreasing susceptibility to UV radiation of coronavirus. The design of the UV radiation in an HVAC system should consider the reported coronavirus susceptibility recognizing that two are dose-response 15 and one is single dose 10 (Table 3) .

Bacteriophage MS2 showed a discrepancy where Tseng and Li 32 found that increasing RH was associated with decreasing susceptibility and Walker and Ko 10 found that increasing RH was associated with increasing susceptibility. Walker and Ko 10 acknowledged that this relationship for bacteriophages (MS2) and animal viruses (adenovirus) was different from that reported previously for bacteria. Other differences between the two studies of bacteriophage MS2 include the susceptibility calculation and suspending medium. Susceptibility was calculated using doseresponse of UV dose and survival fraction by Tseng and Li 32 and using a single dose and survival faction by Walker and Ko. 10 Deionized water was used by Tseng and Li 32 and phosphate buffered saline with 0.01% Tween 80 and Antifoam A was used by Walker and Ko. 10 Three studies examined upper-room UVGI using a room-sized chamber. 12 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 

In the experimental study by McDevitt et al 12 increasing ACH was associated with increasing survival fraction. Two modelling studies confirmed the association of increasing ACH and increasing survival fraction (Table 4) . [43] [44] Increasing ACH is associated with increasing survival fraction because the increased ACH decreases the time that the infectious agent is exposed to UV CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) UV radiation was associated with decreased attack rate which is the number of new cases in population at risk divided by number of persons at risk in population. Zheng et al 48 found that attack rate decreased 87.8% when UVGI was modelled on a cruise ship where the ACH due to UV was 12 ACH.

As early as 1936, UV radiation was associated with decreased influenza transmission in an animal model 49 (Wells, 1936) ( Table 5 ). UV radiation was associated with decreased virus transmission and infection incidence in three of the animal studies. 49-51 Dee et al 52 acknowledged that the lack of effect was likely due to insufficient exposure time to the UV radiation ( Table 5 ).

All of the five human studies 53-57 were from the 1940s and investigated upper-room UVGI (Table 6 ). UV radiation was associated with decreased transmission of respiratory infections in three studies. 53,55-56 In the studies of barracks, Wheeler et al 53 found that high intensity UV treatment was required to decrease transmission. UV radiation was associated with modified spread of transmission, but not prevention of transmission, of measles 54 and chickenpox 57 but not mumps. 57

The risk of bias evaluation for the experimental studies demonstrated three scenarios (Table 7) .

Seventeen studies had low risk of bias for all three domains: selection bias, information bias, confounding. Seven studies had low risk of bias for selection bias and confounding but unclear risk of bias for information bias due to lack of clarity in the description of the UV radiation. Of . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint these seven studies, four were aerosolized virus studies, two were animal studies, and one was a human study. One aerosolized virus study had low risk of bias for selection bias but high risk of bias for information bias and confounding due to lack of calibration of fluorescent bioaerosol count (FBC) and potential for UV radiation to affect fluorescence. Su et al 58 (Table 8) : definition, assumption, validation.

UV inactivation of airborne viruses is governed by the UV dose. The required dose varies depending on the type of virus, capsid structures, and host cell repair mechanisms 59 . Tseng and Li 32 found that dsRNA and dsDNA viruses required a dose that was 2 times higher than their single strand counterparts for 90% inactivation. Walker and Ko 10 came to a similar conclusion when they found that adenovirus was more resistant to inactivation compared to SARS. This resistance is attributed to the double stranded nature of its DNA genome and its ability to shield or consume UV radiation using small proteins concentrated along with viral particles 59 .

External factors such as ventilation and relative humidity also play an important role in UV effectiveness. Relative humidity is uniquely intertwined with UV inactivation. Many studies in this systematic review indicate that relative humidity has a marked, sometimes statistically . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint significant effect on UV inactivation [10] [11] [12] [13] 33 . Further research must be done to ascertain the interaction of relative humidity and UV inactivation in a real-world context with a diverse list of infectious agents. The effect of ventilation on UV effectiveness is similarly complex. Increasing ventilation rate and UVGI are inversely related. At a higher ventilation rate infectious agents are removed from the space at a faster pace. This results in shorter exposure times thereby decreasing the effectiveness of the UVGI system 41 . Dee et al 52 stated that in their experiments UVC provided no reduction in aerosol transmission of PRRSV due to insufficient exposure time. McDevitt et al 12 argued that the combination of increased ventilation and upper room UVC is "more than merely additive". The overall effect of a higher ventilation rate results in an increase in the ACHuv (effective ventilation due to UVC). Li et al 43 came to a similar conclusion that while UV disinfection efficiency decreases when ventilation rate increases, the overall infectious agent removal rate increases. In addition to ventilation rates, airflow patterns can have a meaningful impact on UVGI effectiveness 47 . Well mixed air allows UVGI to be more effective 12, 41, 47 . UVGI systems should be designed specifically for the targeted space. Ventilation rate, airflow patterns, and relative humidity should all be taken into consideration. In spaces with no ventilation systems or where increasing the ventilation rate would not be feasible, UVGI can provide cost effective air disinfection 60 . Spaces with a low outdoor air fraction can also benefit by adopting UVGI 50 . When considering upper room UVGI, lamp placement can have a significant impact on the effectiveness of the inactivation 47 . UV devices set on a wall rather than the corners of the room improves the effectiveness of the UVGI system 43 . Building owners and operators should consult an expert when examining the feasibility of installing UVGI in their space. First et al 41(p328) states that "upper-room UVGI must be approached as a carefully interdependent system with critical interactions among luminaire selection, luminaire placement, and all aspects of a ventilation system."

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint When investigating the effectiveness of UVGI it is important to carefully consider the viral challenge and ensure the experiment mimics natural conditions. Even though Jensen 36(p420) used viral aerosol concentrations that were "many times greater than one would normally expect to encounter under natural conditions", they stated that UVGI used in conjunction with filtration "should kill virtually all viruses". Walker and Ko 10 discussed the importance of aerosolization, sampling, and medium. Some viruses might be inactivated in the act of aerosolization or sampling; a medium with a high protein concentration might protect the targeted virus from inactivation 10 .

UV susceptibility in liquid suspensions cannot be substituted for susceptibility in aerosols. Walker and Ko 10 found that for the viruses they tested, UV susceptibility was higher in aerosols than in is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint respectively, as opposed to large explosive episodes. The effectiveness of UVGI has been known for a long time. The benefits of a well designed UVGI system outweigh the principal and maintenance costs 55 . The time has come for UVGI to be considered as essential as ventilation and filtration.

The results of this systematic review revealed several important findings. First, viruses and bacteriophages were inactivated by UV radiation. Second, increasing UV dose was associated with decreasing survival fraction of viruses and bacteriophages. Third, increasing relative humidity was associated with decreasing susceptibility to UV radiation. Fourth, UV dose and corresponding Studies characterized as in-duct UVGI are designed with mechanically induced air flow with a controlled mean velocity which transports air through an irradiated zone inside a duct. In-duct UVGI lends itself to greater experimental control of the UV dose than the upper-room UVGI, although laminar duct flow, non-uniform velocity profiles, and radiation distribution always lead to some dose variance for in-duct systems. Also, in-duct configurations tend to simulate practical conditions where UVGI is installed in HVAC systems or air purifiers. Non-enclosed systems like upper-room UVGI depend on air circulation to drive the particles to an irradiated zone near the ceiling and does not require a controlled mean velocity: air moves through the UV zone generally . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint due to air currents which are subject to room-scale turbulence. Thus their evaluation becomes more complex, needing to take into account ventilation configurations, air currents and lamp installation locations.

Two main thrusts of the research emerged: (1) research that focused on the effect of UV radiation on aerosolized virus survival and (2) research that considered some or all of the transmission chain from infected host to aerosolized virus to infected target where specific UV radiation configurations or scenarios were evaluated.

Future research must ensure that UV dose and UV design requirements are clearly described within the context of their own study in order to simplify comparison between studies. A common metric used for quantifying the effects of UV on airborne pathogens was the measure of the survival fraction, comparing a quantity before and after the UV intervention. This metric is easy to understand but is highly dependent on the configuration of the system for exposing the aerosol to UV radiation. It shows how a system in its entirety kills or inactivates the pathogen in question, but does not isolate the more fundamental dose-response as the susceptibility (Z) measure, which is the exponent or linear slope on a semi-log Cartesian plot of the UV dose versus survival fraction.

An encouraging and important trend evident in this body of work is the general evolution toward more rigorous control of experimental conditions which lend themselves to clear quantification of dose-response. Knowing that not all UV sources are alike, and that flux divergence, reflection, air velocity profile and lack of turbulence can lead to non-uniform UV radiation of aerosols allows researchers to focus on comparing susceptibility (Z) values, a single parameter. In some cases, the UV radiation and aerosolized virus studies were able to provide the mechanistic quantitative . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint measure of susceptibility (Z) which is useful for cross study comparisons and is an important input parameter in modelling studies.

In addition to research examining the effect of UV radiation on aerosolized virus survival, the other main focus was research considering some or all of the transmission chain from infected host to aerosolized virus to infected target under specific UV radiation conditions. These results tend to be somewhat anecdotal because there is no standard test for the full transmission chain between two or more people sharing breathing space in the built environment. This is a common challenge in many fields of research. There would be value if the community moved to a more standardized test case configuration such as a standard room and ventilation system configuration, so that discrepancies attributable to factors such as geometry and flow field could be eliminated. The ASHRAE 185.1 standard states "Test standards form the foundation for air-cleaner selection in the ventilation industry. U.S. Environmental Protection Agency (USEPA) literature states that the most important need in the area of ultraviolet germicidal irradiation (UVGI) is industry standards to rate installed devices." 62(p2) All human studies of upper-room UVGI that we identified were field studies conducted between 1945-1949. In a historical review, Reed 63 indicates that UVGI fell out of favour after the 1940s due to inconsistent UV effects on measles transmission in schools, which were later attributed to measles exposure outside of schools. Additionally, no human studies on virus transmission and UV radiation were found after the 1940s which is a time period before vaccinations for measles, 64 chickenpox, 65 and mumps. 66 Reed 63 attributes more recent attention on UVGI to bacteria, tuberculosis, and viruses, influenza and SARS. In the current review, recent studies of UVGI are modelling studies in hospital settings from 2006 to 2020 in which influenza (2011) and coronavirus . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint (2020) were considered. More field studies of upper-room UVGI are warranted to advance our understanding of its applicability.

The design of the UVGI for implementation in a building environment, whether in-duct or upperroom, should consider which virus is targeted. An additional consideration is how the susceptibility of that virus is affected by changes in relative humidity (Figure 2 ). If more than one infectious agent is targeted, then a range of susceptibilities should be considered. Jaynes et al 51 designed their UVGI system to target one bacteria and two viruses ( Table 5 ). Practical application of UVGI systems should take into account lessons learned from modelling studies; i.e. UV dose produced by the UV system may be affected by the airflow pattern, ACH and UV device location. [43] [44] [45] 47 Generally, in-duct UVGI can be used to prevent in-building transmission when ventilation supply air has some fraction which is recycled, a practice which is a common means to reduce heating/cooling costs. Also, in-duct UVGI could be incorporated into a ducted air purification system which exhausts back into the source space directly and might be a tool used to remove airborne virus in a space at a rate higher than could be achieved by the building air handling system alone. In both of these cases the survival of pathogen after UV exposure would need to be extremely low for them to be effective (though in practice in-duct systems would generally operate in series with aerosol removal by filtration). A consensus on an acceptable standard for virus reduction due to UV treatment might be helpful. Experimental studies may benefit from using the ASHRAE Standard 52.2 test duct which has been used in recent filtration studies. [67] [68] Simply seeing more studies recognize that UV radiation dose is dependent on the radiant flux, which is often not uniform, shows a positive trend. For ducted systems, flow that is well-mixed and close to uniform in velocity profile is helpful in evenly dosing aerosols with UV.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) conventional UV can be carcinogenic and cataractogenic 34,44 and a health hazard when exposed directly. 15 Far-UVC light has been posited as an option for UV radiation as, since far-UVC light has a lower range "of less than a few micrometers, and thus it cannot reach living human cells in the skin or eyes," the range is still greater than that of viruses, allowing UVC light to "penetrate and kill them. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint UVGI need to take into account the various factors that exist within ventilated indoor spaces that may modify UV effectiveness, including humidity, airflow pattern, air changes per hour, and UV device location. Research is needed to provide evidence of the effect of UV radiation along the chain of transmission in non-simulated "real life" settings.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint Unpublished, not peerreviewed CFD = computational fluid dynamics; HVAC = heating, ventilation, and air conditioning; MERV = minimum efficiency reporting value; UVGI = ultraviolet germicidal irradiation . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. For 20 days evaluated, 12 days had significantly lower FBC in UVGI rooms compared with non-UVGI rooms, 6 days had -UV radiation associated with reduction of fluorescent bioaerosol counts . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 14, 2021. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Data reported as (95% confidence interval) or ± standard deviation. ssRNA = single-stranded ribonucleic acid; ssDNA = single-stranded deoxyribonucleic acid; dsRNA = double-stranded ribonucleic acid; dsDNA = double-stranded deoxyribonucleic acid . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) CFD = computational fluid dynamics SEIR = susceptible-exposed-infected-recovered . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) * inconsistencies in units for dose appear to be typos in the original paper, we assumed these to be 420, 840 and 1260 J/cm 2 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint UV radiation associated with decreased transmission.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) -UV radiation associated with modified spread of transmission for chickenpox.

-UV radiation not associated with modified spread of transmission for mumps.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint Figure 1 . Flow of studies through the selection process (note: search was conducted for all HVAC design features but only studies of UV radiation are included in this manuscript).

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint Figure 2 . UV radiation susceptibility (Z) and relative humidity (RH) Each colour represents one study. Reported RH ranges shown as average RH values. ~C utler at al 33 reported that susceptibility (Z) was significantly lower at ≥80%RH (shown at 80%RH) compared with 25%-79%RH (shown at 52%RH). *Walker and Ko 10 calculated susceptibility from a single dose and corresponding survival fraction, rather than dose-response of UV dose and survival fraction.

^ Lin et al 35 calculated susceptibility from a single dose and corresponding survival fraction for "fast decay" and from the dose-response of UV dose and survival fraction for "slow decay".

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 14, 2021. ; https://doi.org/10.1101/2021.10.12.21264904 doi: medRxiv preprint

speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-oncovid

Transmissibility and transmission of respiratory viruses

A systematic review of possible airborne transmission of the COVID-19 virus (SARS-CoV-2) in the indoor air environment

Bidirectional association between COVID-19 and the environment: a systematic review

Role of mechanical ventilation in the airborne transmission of infectious agents in buildings

Role of ventilation in airborne transmission of infectious agents in the build environment -a multidisciplinary systematic review

Role of viral bioaerosols in nosocomial infections and measures for prevention and control

The Application of Ultraviolet Germicidal Irradiation to Control Transmission of Airborne Disease: Bioterrorism Countermeasure

The use of ultraviolet germicidal irradiation (UVGI) in disinfection of airborne bacteria and rhino-viruses

Effect of ultraviolet germicidal irradiation on viral aerosols

Characterization of UVC light sensitivity of vaccinia virus

Inactivation of Poxviruses by Upper-Room UVC Light in a Simulated Hospital Room Environment

Aerosol susceptibility of influenza virus to UV-C light

A Critical Review on Ultraviolet Disinfection Systems against COVID-19 Outbreak: Applicability, Validation, and Safety Considerations

Effect of ultra-violet irradiation of classrooms on spread of mumps and chickenpox in large rural central schools -A progress report

A case study of upper-room UVGI in densely-occupied elementary classrooms by real-time fluorescent bioaerosol measurements

Inactivation of feline calicivirus and adenovirus type 40 by UV radiation

The application of ultraviolet germicidal irradiation to control transmission of airborne disease: bioterrorism countermeasure

Germicidal Efficacy and Mammalian Skin Safety of 222-nm UV Light

Method of Testing UV-C Lights for Use in Air-Handling Units or Air Ducts to Inactivate Airborne Microorganisms

The history of ultraviolet germicidal irradiation for air disinfection

Chickenpox vaccination: What everyone should know

Mumps vaccination

Study of viral filtration performance of residential HVAC filters

Challenge of mechanical and antimicrobial filters against infectious phages artificially agglomerated with inorganic dust with a known particle-size distribution