The COVID-19 or SARS-CoV-2 outbreak, first detected in Wuhan, China, was characterized as a pandemic by the World Health Organization (WHO) in March 2020 [1]. Almost 2 years later (early February 2022), the WHO reported over 383 million cases of the disease (COVID-19) caused by the virus (SARS-CoV-2), with over 5.6 million deaths worldwide [2]. Early in the pandemic, there were conflicting views and debate about the routes of transmission [3-6]. Several recent reviews of the scientific literature have identified evidence indicating airborne transmission, which could be particularly problematic in confined and crowded indoor spaces [7-9]. Public health recommendations acknowledge airborne transmission as important and advise to maximize ventilation; ensure proper maintenance and functioning of heating, ventilation, and air-conditioning (HVAC) systems; and increase the use of fresh air where possible [10].

Airborne transmission occurs as a result of bioaerosols (biological particles suspended in air) staying aloft longer because of their small size and, therefore, traveling further because of air currents [3]. Several possible mechanisms of airborne coronavirus transmission exist, including 1) bioaerosol generation by infectious persons through coughing, sneezing, breathing, and talking, which remain airborne for a period of hours to days; 2) short- to long-range transport through HVAC systems and subsequent inhalation of bioaerosols by other people; and 3) airborne transport of bioaerosols to surfaces (or the contamination of surfaces by physical contact), followed by resuspension, inhalation, or contact with surfaces [11,12].

Previous research demonstrated that infectious airborne bioaerosols spread to other spaces via HVAC systems [12,13]. Multiple features within HVAC systems may influence transmission, including ventilation (eg, ventilation rate, air changes per hour, airflow pattern, and pressurization), filtration (eg, minimum efficiency reporting value rating, filter age, and extent of use), UV radiation (eg, UV power and UV dose), and humidity [12]. Understanding the influences of HVAC systems on airborne transmission in the built environment is critical for building scientists to develop effective engineering control strategies to protect the occupant’s health and well-being and affect timely public health policies. Previous systematic reviews provided a starting point for understanding what is already known from the scientific literature about HVAC systems and the airborne transmission of viruses. A comprehensive synthesis of previous systematic reviews can help identify knowledge gaps, helping to guide and prioritize future primary research. Therefore, we conducted an overview of reviews to identify and synthesize previous systematic reviews on this topic.

A research librarian (GMT) conducted searches in Ovid MEDLINE and Compendex from inception to June 2020, using concepts related to viruses, transmission, and HVAC. The search was updated in January 2021. The search strategies are presented in Multimedia Appendix 1. The unfiltered search strategies were peer reviewed by 2 librarians (TL and AH), and the filter for systematic reviews in Ovid MEDLINE was provided by a third librarian (LD). The unfiltered search strategies were part of a larger systematic review project that was registered [15], and its protocol is publicly available [16]. The reference lists of the included reviews were screened to identify any other relevant reviews. Conference abstracts and preprints retrieved through the searches were screened to determine whether a full peer-reviewed manuscript was published. The references were managed in EndNote; duplicate records were removed before screening.

Two reviewers (GMT and LH) independently screened the titles and abstracts of all the citations retrieved from the electronic searches and other sources. Studies were classified as yes, no, or maybe. The first stage of screening was completed in Covidence. We retrieved the full text of all the studies classified as yes or maybe. The same reviewers independently applied the inclusion and exclusion criteria (Multimedia Appendix 2 [16]) to each full-text document and classified the studies as included or excluded. Discrepancies were resolved through discussion between the 2 reviewers. The reasons for excluding studies at the full-text stage were documented.

The inclusion and exclusion criteria are detailed in Multimedia Appendix 2. We planned to include systematic reviews published in English that searched for and included primary research studies examining the effects of HVAC design features on the transmission of viruses. The HVAC features of interest were mechanical ventilation (ventilation rate, air change, air exchange, and airflow), filtration (air filtration, filter type, minimum efficiency reporting value rating, filter age and use, pressure drop, holding capacity, replacement, and change frequency), UV germicidal irradiation (power, dose, uniformity of dose, flow rate, bioaerosol inactivation efficiency, and location), and humidity or relative humidity (RH). Inclusion was staged in 2 ways. Our primary interest was viruses, and we excluded those reviews that were not specific to virus. We were initially interested in systematic reviews defined by the international Cochrane organization as reviews that use a predefined, systematic approach and follow standard approaches to search the literature, select studies for inclusion, assess the methodological quality of the included studies, and extract, synthesize, and analyze data from the included studies. As we found few systematic reviews meeting these criteria, we included review articles that satisfied specific requirements for methodological approach and objective. For methodological approach requirements, the authors had to search ≥2 databases, describe inclusion and exclusion criteria, and describe a process for study selection. For objective requirements, the objective of the review had to be related to one of the HVAC design features, namely ventilation, filtration, UV radiation, or humidity.

The methodological quality of the included reviews was assessed using AMSTAR2 [17]. AMSTAR2 is a valid and reliable tool containing 16 items about the methodological conduct of a systematic review [18]. Two authors (GMT and LH) independently assessed the included reviews. Discrepancies were resolved through discussion.

The following information was extracted from each review: citation information (eg, authors, year of publication, and country of corresponding author), objectives, search strategy, inclusion and exclusion criteria, settings, population characteristics (as applicable), agent studied (eg, type of virus and bioaerosol), HVAC design features studied, number and characteristics of studies relevant to this overview’s research question, results (as reported by the review authors), and review authors’ conclusions relevant to this overview’s research question. Our primary outcome was the quantitative measure of the association between HVAC design features and virus transmission; however, we extracted any results reported by the review authors that were relevant to our research question. One reviewer (LH) extracted data using a predefined form. A second reviewer (EK) verified the data. Discrepancies were resolved through discussion and by referring to the relevant publication.

We anticipated that the included reviews would not have conducted meta-analyses. We planned to present the results in tabular and narrative forms. Tables were created describing the reviews, their results (including any quantitative data of the associations between HVAC features and virus transmission or proxy outcomes) and conclusions, and their methodological quality. A narrative summary of the findings of each review has been provided. We only summarize review findings that were relevant to our research question; for example, if the review included studies of ventilation, humidity, etc, in the outdoor and indoor environments, we only report on studies specific to the indoor (built) environment.

The search retrieved 361 citations, of which 45 (12.5%) were considered potentially relevant and 7 (2%) met the inclusion criteria (Figure 1). Tables 1 and 2 provide summaries of the included reviews. The reviews varied somewhat in their objectives (eg, investigate mechanical ventilation, ventilation rates, airflow patterns, effects of humidity, or stability of bioaerosols containing coronaviruses), agents (eg, coronaviruses or influenza viruses), and settings (eg, built environment, health care settings, or public ground transportation). The reviews were published between 2007 and 2021 (median year 2020) and included a total of 47 unique virus studies published between 1961 and 2020 (median year 2005) that were relevant to our research question (median 4 studies per review including shared references; Table 3; Multimedia Appendix 3 [13,19-71]). The reasons for excluding studies at the full-text stage were documented (Multimedia Appendix 4 [7-9,11,12,24,72-103]).

Li et al [13] examined the role of ventilation (specifically, ventilation rates and airflow patterns) in the airborne transmission of infectious agents in indoor settings. The authors included 40 English-language studies overall, with 16 (40%) specific to viruses reported between 1962 and 2005 (median year 1985/1996). Of the 16 studies, 3 (19%) included multiple papers (Table 3), which increased the total count to 21. Of these 21 studies, 16 (76%) studies were epidemiological, 4 (19%) involved other observational designs, and 1 (5%) was experimental. Of the 21 studies, 3 (14%) studies had limited and 4 (19%) had no investigation of ventilation rates or airflow. Studies involved a variety of settings: hospitals, hospital wards, or health clinics (9/21, 43%); aircrafts (3/21, 14%); nursing homes (3/21, 14%); schools (2/21, 10%); high-rise apartments (2/21, 10%); an office (1/21, 5%); and an animal cage (1/21, 5%). The viral agents included severe acute respiratory syndrome coronavirus 1 (SARS-CoV-1; 7/21, 33%), influenza (5/21, 24%), measles (4/21, 19%), chicken pox (2/21, 10%), rhinovirus (1/21, 5%), common cold (1/21, 5%), and smallpox (1/21, 5%). Overall quality was assessed as good for 12 (57%) studies, average for 5 (24%) studies, and unsatisfactory for 4 (19%) studies. The researchers convened a panel of experts in medicine, public health, and engineering. They used a modified Delphi approach with a final consensus meeting to rate the “evidentiary threshold” to support their hypothesis, that is, the direct contribution of ventilation to airborne transmission. Among the virus studies, 8 (38%) were rated as conclusive, 8 (38%) were partly conclusive, and 5 (19%) were nonconclusive. Among the 8 conclusive studies, 2 (25%) examined ventilation rates and showed higher rates of infection for influenza with lower ventilation rates and 6 (75%) demonstrated an association between airflow patterns and the transmission of measles (pediatric office suite), chicken pox (hospital), smallpox (hospital), and SARS-CoV-1 (hospital). In all the studies, the bioaerosols traveled a “considerable distance,” which the reviewers noted, “seemed to be related to building design” [13] (eg, placement of heating radiator, room pressure, and functional status of return air outlet). None of the virus studies provided data to support “specification and quantification of the minimum ventilation requirements” [13].

Luongo et al [19] examined evidence from epidemiological studies for the association of mechanical ventilation (at least one HVAC parameter) with the airborne transmission of infectious agents in buildings. Although the authors included 13 English-language studies, 3 (23%) were specific to viruses; all the studies were observational and were reported between 1996 and 2011 (median year 2004). One of the studies had 2 papers, which increased the total count to 4 (Table 3). All 4 virus studies were also included in Li et al [13]. The settings included nursing homes (2/4, 50%), an office building (1/4, 25%), and a hospital (1/4, 25%). The viruses represented in the studies included influenza (2/4, 50%), SARS-CoV-1 (1/4, 25%), and rhinovirus (1/4, 25%). The review authors did not assess methodological quality but provided a narrative commentary on the strengths and limitations of each study. Of the 4 studies, 2 (50%) found an association between virus incidence rates, self-reported incidence rates, and the risk of exposure with HVAC design features. In a retrospective cohort study of a SARS-CoV-1 outbreak in a hospital, the authors measured ventilation rates and found that “proximity to index patient associated with transmission” [19]. The authors of the second study blindly adjusted outdoor air supply dampers in 3 office buildings and found a significant positive association between average CO₂ concentration greater than 100 ppm above background and the frequency of rhinovirus detection in air filters. The third study found a lower incidence of influenza in newer nursing homes that had 100% outside air delivery (compared with older homes with 30%-70% recirculated air) and filtered room supply (compared with no filtration) during 1 season; however, data collected over 5 subsequent years, reported in the fourth study, found no clear association. None of the studies quantified the minimum ventilation requirements.

Derby et al [20] conducted a literature review to assess the effects of low humidity (≤40% RH) on comfort, health, and indoor environmental quality. Although the review included approximately 70 papers, 9 (13%) papers examined the effects of humidity on the viability or transmission of airborne viruses. Of these 9 studies, 7 (78%) were experimental (involving laboratory testing), 1 (11%) was a reanalysis of the data from one of the experimental studies, and 1 (11%) study involved modeling. Most studies focused on influenza, with 1 study each examining Columbia SK viruses, murine norovirus, and multiple viruses (influenza, vaccinia, Venezuelan equine encephalomyelitis, and poliomyelitis). Most studies examined a wide range of RH, from approximately 5% to 25% RH at the lower range to 75% to 100% RH at the upper range. The absolute humidity was approximately ≤25 g/m³ in all 9 studies except 1 (11%) (which ranged from 25 to 125 g/m³). The review authors did not assess the methodological quality of the included studies. In terms of virus viability, 4 studies (44%) showed a reduction in midrange RH (ie, approximately 50% RH). The review authors further noted that 5 studies (56%) showed that “virus survival exhibited a canonical dip between 40 and 80% RH” [20] and that in almost all the cases, the decline in survival was correlated with increased length of exposure. In total, 3 studies (33%) examined influenza transmission. One study showed reduced influenza transmission among guinea pigs at 50% RH versus 20% to 35% RH; however, the same pattern was not found when the researchers analyzed the data based on absolute humidity. A second study examined transmission via coughing using manikins and found 5 times more infectious virus at 7% to 23% RH than at >43% RH. A modeling study of influenza virus transmission via coughing showed that the infectious virus concentration was 2.4 times more at 10% RH than at 90% RH after 10 minutes, and the ratio increased over time. They also demonstrated that the effect of humidity is related to the particle size: the settling of larger particles and inactivation of smaller particles (<5 µm) with greater humidity. They concluded that the inactivation resulting from high RH coupled with ventilation was important to remove smaller particles.

Chirico et al [21] conducted a rapid review (streamlined systematic review methods) to examine the potential role of air-conditioning (HVAC) systems in “outbreaks of coronaviruses (SARS-CoV-1, MERS-CoV, SARS-CoV-2) in indoor environments” [21]. The authors identified 14 studies published between 2003 and 2020 (n=11, 79% peer-reviewed studies and n=3, 21.4% preprints all concerning SARS-CoV-2); the studies investigated outbreaks in Hong Kong (n=7, 50%), South Korea (n=1, 7%), Japan (n=3, 21%), and China (n=3, 21%). Of 14 studies, 7 studies (50%) examined 2 outbreaks associated with SARS-CoV-1: 5 (71%) studies examined outbreaks (different areas or groups of individuals) within the same hospital, and 2 (29%) studies investigated an outbreak in the same private high-rise housing estate. Of the 7 SARS-CoV-1 studies from Chirico et al [21], 5 (71%) are shared references with Li et al [13] (Table 3). The review authors indicated that 6 of 7 SARS-CoV-1 (86%) studies indirectly demonstrated a role for the HVAC system (through epidemiological data, spatiotemporal patterns of infection, or modeling). Of 14 studies, 1 (7%) study investigated an outbreak of Middle East respiratory syndrome coronavirus (MERS-CoV) in a hospital setting and demonstrated the contamination of the HVAC system through environmental sampling [22]. Of 14 studies, 6 (43%) studies investigated outbreaks of SARS-CoV-2: 1 (17%) study examined 318 outbreaks in 120 cities in China, including community and workplace settings; 3 (50%) studies examined an outbreak on a ship in Japan; and 2 (33%) studies examined the same outbreak in a restaurant. A total of 3 (50%) observational studies suspected a role for the HVAC system, 2 (23%) studies (both of ship outbreak) did not find evidence of a role for HVAC based on the spatiotemporal distribution of cases, and 1 (17%) study (of restaurant outbreak) supported a role for HVAC by computer simulation. The review authors indicated that they were not able to appropriately evaluate the quality of the included studies. The review authors concluded that there is sufficient evidence from SARS-CoV-1 and MERS-CoV studies demonstrating a role for HVAC in the airborne transmission of the viruses; however, there was not sufficient evidence that HVAC systems play an important role in the case of SARS-CoV-2. Although there is a lack of evidence for SARS-CoV-2, there was no evidence of no role.

Zhen et al [23] conducted a rapid review of “the role of public ground transport in COVID-19 transmission” and “interventions that may reduce transmission” [23]. The authors searched for studies published since 2000 and identified 4 relevant studies, published between 2007 and 2016, namely 1 (25%) systematic review, 1 (25%) case-control study, and 2 (50%) modeling studies. The systematic review by Browne et al [24] identified 41 studies examining the risk of transmission of Influenza A (H1N1/09) (n=29, 71% studies), SARS-CoV (n=5, 12% studies), both influenza and SARS-CoV (n=2, 5% studies), MERS-CoV (n=2, 5% studies), or unspecified viruses (n=3, 7% studies) related to sea (n=6, 15% studies), ground (n=6, 15% studies), or air (n=29, 71% studies) transport. Zhen et al [23] summarized results from 4 quantitative studies included in Browne et al [24] and concluded that the “use of public transport increased the risk of influenza transmission” [23]. Zhen et al [23] identified a multicenter case-control study that showed a lower probability of Influenza A (H1N1/09) diagnosis with public transport use (metro, bus, tram, or local train) and no association with diagnosis and the use of trains, airplanes, or taxis. The case-control study was assessed as having a moderate risk of bias by Zhen et al [23]; the risk of bias was not reported for the other 3 studies (75%). Zhen et al [23] also identified 2 modeling studies: one estimated the reproduction number for influenza infection in a train, and the other tested simulations to predict influenza infection probability for 4 bus ventilation systems. The first modeling study showed that masks could decrease the reproduction number, resulting in a lower risk of disease transmission, with high-efficiency particulate air masks being more effective than surgical masks. Furthermore, doubling the ventilation rate reduced the risk, similar to the use of high-efficiency particulate air masks, and was considered more feasible and cost-effective. The second modeling study showed that influenza transmission risk can be reduced when the infected passenger is positioned closer to the exhaust opening and with high-efficiency filtration in the case where nonrecirculated air cannot be provided. Given the limited number of research studies, Zhen et al [23] also identified and discussed national and international guidance documents, for example, those published by WHO [104-110]. Although general recommendations have been made to reduce risk (eg, minimizing the use of public transport, environmental controls, respiratory etiquette, hand hygiene, and mask use), there is no indication of the empirical evidence specific to these measures, in particular mechanical ventilation.

da Silva et al [25] conducted a systematic review to discuss “the viability/stability of aerosols containing SARS-CoV and MERS-CoV viruses” with an intent “to provide information on potential mitigation strategies for SARS-CoV-2 airborne transmission” [25]. The review authors identified 11 studies. Of these 11 studies, 8 (73%) studies examined the viability of coronaviruses in air samples, but the review authors did not describe the relationship with HVAC features, including 1 (13%) MERS-CoV study [22], which was described earlier by Chirico et al [21]. The remaining 3 (27%) studies were laboratory-based experimental studies of coronaviruses. In one MERS-CoV study, the virus was aerosolized at 20 °C with 40% or 70% RH, showing decreased stability at 70% RH compared with 40% RH [26]. The other MERS-CoV study examined virus inactivation under 2 conditions [27]: common office environment (25 °C and 79% RH) and Middle Eastern region climate (38 °C and 24% RH). In the simulated office environment, “the virus demonstrated high robustness and strong capability to survive with about 63.5% of viruses remaining infectious 60 min after aerosolisation. Virus decay was much stronger for hot and dry air scenario with only 4.7% survival over 60 min procedure” [25]. One of the studies showed an aerosol survival time of 3 hours for SARS-CoV-2 [28]. The review authors did not assess the methodological quality of the included studies; however, they commented on some limitations. The review authors concluded that “higher temperatures and high relative humidity can have an effect on SARS-CoV-2 viability in the environment as reported in previous studies to this date” [25]. However, their conclusions were based on studies of both indoor and outdoor environments.

Noorimotlagh et al [29] performed a systematic review of SARS-CoV-2 literature “to collect all available studies concerning inactivation methods, environmental survival, and control and prevention strategies” [29]. Although 42 studies were identified, 4 provided information on temperature and humidity in the built environment, investigating MERS-CoV (n=2, 50%), SARS-CoV-1 and SARS-CoV-2 (n=1, 25%), and Phi6 (n=1, 25%), which is a bacteriophage used as a surrogate for viruses. All of them were laboratory-based experimental studies. The review authors did not assess or comment on the methodological quality of the included studies. The aerosolized MERS-CoV data from van Doremalen et al [26] were reported by da Silva et al [25] although not extracted by Noorimotlagh et al [29]. A MERS-CoV study found that the virus had robustness and strong capability to survive at 25 °C and 79% RH [27]; this was a shared reference with da Silva et al [25]. Although da Silva et al [25] reported the SARS-CoV-2 survival time from van Doremalen et al [28], Noorimotlagh et al [29] further clarified that the aerosol survival time of 3 hours for SARS-CoV-2 was at 40% RH and 21–23 °C and that the stability of SARS-CoV-2 was similar to that of SARS-CoV-1 [28]. The Phi6 study showed the highest virus survival at >85% RH and <60% RH with a significant decrease between 60% and 85% RH [30]. At a fixed humidity of 75% RH, infectivity decreased by 2 orders of magnitude between 19 and 25 °C [30]. The review authors concluded that “temperature and relative humidity are important factors in the survival of SARS-CoV-2” [29] and that “disease transmission via droplets is inhibited by increasing both temperature and RH in buildings” [29]. A review recommendation was “proper ventilation of the buildings in time of aerosol generating” [29]; however, studies of ventilation were not reviewed.

The network of the 7 included reviews and their 47 references relevant to this overview was created using Palladio (Figure 2). Overall, 12 references were shared among the 7 included reviews. However, the network clearly demonstrates that Derby et al [20] and Zhen et al [23] shared no references with the 5 other reviews. In actuality, the 12 references were shared between 5 reviews (Figure 2). da Silva et al [25] and Noorimotlagh et al [29] shared 3 references regarding experimental studies of MERS-CoV and SARS-CoV-2 [26-28] (Table 3). In addition, da Silva et al [110] shared 1 reference on MERS-CoV isolation wards [22] with Chirico et al [21]. Three reviews, Li et al [13], Luongo et al [19], and Chirico et al [21], shared a reference on SARS-CoV-1 in hospital wards [31]. Li et al [13] and Chirico et al [21] shared other related references on SARS-CoV-1 in hospital wards [32,33]. Similarly, Li et al [13] and Luongo et al [19] shared studies on influenza in nursing homes [34,35] and rhinovirus in offices [36]. Li et al [13] and Chirico et al [21] shared 2 other references regarding SARS-CoV-1 in high-rise apartment complexes [37,38]. Not only were the 12 references shared only between 5 reviews, 8 of these references were shared with 1 2007 review by Li et al [13], and the remaining 4 references were shared with 1 2021 review by da Silva et al [25] (Figure 2). Although 35 of the 47 references were not shared, the 12 shared references were captured by the earliest review (2007; [13]) and one of the latest reviews (2021; [25]).

Table 4 provides the assessments of the methodological quality of the reviews based on AMSTAR2. In total, 3 (43%) papers described themselves as systematic reviews, 2 (29%) were rapid reviews, 1 (14%) was a broad literature survey, and 1 (14%) was described simply as a review. The majority provided detailed research questions, explained study designs considered for inclusion, used a comprehensive search strategy, described the included studies, discussed the heterogeneity of results, and reported potential conflicts of interest. None or few reviews provided an a priori protocol, performed study selection and data extraction in duplicate, provided a list of excluded studies, conducted risk of bias assessments of individual studies, or reported on the sources of funding for the included studies. None of the reviews conducted a meta-analysis; all of them provided a narrative synthesis of the results and observations across the included studies. A previous review [19] spoke about the need for more well-designed studies (including representative sampling and clear and consistent measurement methods and reporting of data) with the goal of using meta-analysis to integrate the results.

This comprehensive overview of reviews provides a map of the existing synthesized evidence on the role of HVAC in airborne virus transmission. The earliest review by Li et al [13] published in 2007 found evidence of an association between ventilation rates and airflow patterns in buildings and the transmission of viral diseases. Li et al [13] found no studies that provided minimum ventilation requirements to prevent the spread of viral diseases; however, they found 1 study showing that tuberculin conversion was significantly associated with ventilation rates of <2 air changes per hour in general patient rooms [111]. Published in 2007 shortly after the 2003 SARS-CoV-1 epidemic, Li et al [13] called for a “multidisciplinary research culture” to study outbreaks, as well as smaller-scale transmission occurrences, for filling the gap with respect to quantifying minimum ventilation standards in both clinical and nonclinical settings. A subsequent review by Luongo et al [19] published almost 10 years later, in 2016, included a subset of 4 of the virus studies identified by Li et al [13], with similar conclusions about the possible association between ventilation features (low outdoor air supply and imbalance in supply and exhaust airflow rates) and airborne virus transmission. Luongo et al [19] also pointed out the lack of data to quantify how mechanical ventilation may affect airborne transmission and the need for more well-designed multidisciplinary epidemiological studies. More recently, in response to the current COVID-19 pandemic, Chirico et al [21] examined HVAC systems and their role in the airborne transmission of coronaviruses; they concluded that there was sufficient evidence demonstrating an association for SARS-CoV-1 and MERS-CoV, whereas there was a lack of evidence for SARS-CoV-2. Derby et al [20] specifically examined the role of humidity in relation to indoor air quality; the evidence they identified was specific to influenza and showed that virus survival was lowest between 40% and 80% RH and that survival time decreased with the length of exposure to humidity. One of the studies from Noorimotlagh et al [29] indicated that aerosolized SARS-CoV-2 can survive for 3 hours at 40% RH and 21 to 23 °C [28]. In another recent review published in 2021, da Silva et al [25] examined mitigation strategies and found 2 studies demonstrating that coronavirus transmission decreased with increasing both temperature and RH in buildings. A recent review (2020) by Zhen et al [23] examined interventions to reduce virus transmission in public ground transportation; 2 modeling studies showed ventilation and filtration to be effective.

Although there is an extensive body of literature examining HVAC and its role in airborne virus transmission, there is a lack of empirical evidence to quantify the minimum standards for HVAC design features in the built environment. Previous reviews have discussed this gap, stressed the need for methodologically rigorous epidemiological studies involving multiple disciplines (eg, engineering, medicine, epidemiology, and public health), and discussed considerations for future research, including the specificity of the virus, its construction and envelope composition, the infectious dose, and the size of the particle containing the virus. The review authors have called for standardizing experimental conditions, measurements, terminologies, and reporting as well as simulating real-world conditions [19,20,25]. An important consideration in designing rigorous studies is controlling for confounding factors. HVAC systems operate in a complex environment; for example, Derby et al [20] noted several confounding variables to be considered when interpreting their findings on humidity and temperature including “variation in air exchange rate, length of organism exposure, variation in the biological structure and routes of entry, variation of pathogen survival on different fomites, and variances in human host response” [20]. They further noted that the number and complexity of the variables to consider “greatly increases the test matrices required” [20] to build a comprehensive evidence base. Studies have also demonstrated the importance of the positioning of the infected person relative to HVAC features and other occupants, mobility patterns and activities (eg, type and intensity of respiratory activity) of the occupants, time spent within a space, occupancy, and occupant density. Despite the specification of airflow parameters, the flow of air in occupied spaces is almost always turbulent (vs laminar) such that particles “are constantly mixing and moving in varied ways across a space,” making assessments and predictions challenging [112]. Finally, research results need to be interpreted in light of the technological differences in the HVAC systems around the world [21]. Engineers have developed sophisticated methods (through modeling, computational fluid dynamics, etc) that allow for the isolation of features and control for confounding variables. However, these studies rely on many assumptions that may not hold in real-world settings or are specific to an assumed building design or configuration. In addition, these studies may isolate 1 component in the chain of transmission, which does not necessarily equate to the actual disease (eg, the detection of viral particles vs infectivity vs disease outcomes) [19,20,25]. The results from modeling studies need to be considered alongside epidemiological studies. Previous reviews have highlighted many challenges with studying outbreaks: Li et al [13] mentioned that the “most inherent limitation in almost all existing investigations is due to the rapid disappearance of airborne evidence of infection, once the infectious period is over” [13]. They proposed as a solution “contemporaneous air-sampling and environmental measurements” [13] in locations during a patient’s illness, which could be extended to locations of high use or occupancy during a pandemic or seasonal epidemics.

The strengths of this study include its comprehensiveness and the use of methods to avoid bias, such as the prespecification of inclusion and exclusion criteria and involvement of at least two reviewers at all stages. The main limitation stems from the limits of the included reviews. We initially intended to include only systematic reviews that met internationally recognized definitions and methodological expectations. However, we relaxed our criteria given that many reviews did not meet this standard. Although most reviews prespecified their research question and conducted a comprehensive search, few conducted study selection and data extraction in duplicate as recommended to avoid bias, and very few assessed the methodological quality or risk of bias of the included studies, which is key to determining the validity and certainty of the available evidence. We also did not find reviews of all HVAC design features; for example, none of the included reviews examined UV germicidal irradiation (although a recent narrative review has been published in the context of COVID-19 [72]), and only a small number of studies across the reviews examined filtration.

The findings of this overview have several implications for public health measures to mitigate the spread of viral transmission in buildings. First, ventilation rates and airflow patterns have been shown to be associated with virus transmission. Second, humidity and temperature are associated with virus survival. Third, filtration can be effective in removing pathogens if the filter rating is commensurate with the size of the particles of interest [19]. The reviews have also mentioned the importance of regular maintenance of HVAC systems and features to ensure optimal functioning. Across the reviews, there was a clearly stated need for more methodologically rigorous interdisciplinary research with a specific focus on quantifying the minimum specifications for HVAC features. Although one of the reviews did not find sufficient evidence of association between HVAC and airborne transmission specific to SARS-CoV-2, the authors did advise (based on evidence for MERS-CoV and SARS-CoV-1) that attention be given to the design and management of HVAC systems as a precautionary measure until further evidence indicates otherwise [21].

Airborne transmission is now recognized as a route of transmission for different viruses, including coronaviruses, specifically SARS-CoV-2, which has been the source of immense global impacts in terms of morbidity, mortality, and the peripheral effects of pandemic restrictions. HVAC systems and their specific features have the potential to mitigate transmission in built environments: there is evidence that ventilation rates, airflow patterns, humidity, temperature, and filtration can influence virus transmission. Enhancing HVAC systems in built environments (including schools, office buildings, commercial spaces, recreation centers, and transport vehicles) could have important implications for the current pandemic as well as seasonal epidemics and other diseases and impacts that are associated with general indoor air quality. These measures will be of utmost relevance to countries that experience cooler climates and where people spend an inordinate amount of time (80%-90%) indoors. Moreover, mitigation strategies that do not rely on human behavior and result in other (eg, social) consequences will be more sustainable [21]. This overview synthesized 7 previous reviews that included 47 studies examining HVAC design features and their effects on the airborne transmission of viruses, serving as a starting point for future systematic reviews and identifying priorities for primary research.