Abstract
Objective Given possible airborne transmission of SARS-CoV-2, an overview of reviews was conducted to understand what is already known from the scientific literature about how virus transmission may be affected by heating, ventilation, and air-conditioning (HVAC) design features in the built environment.
Methods Ovid MEDLINE and Compendex were searched from inception to January 2021. Two reviewers independently screened titles and abstracts and full text of potentially relevant reviews, using a priori inclusion criteria. Inclusion criteria were systematic reviews examining effects of HVAC design features on virus transmission. Two reviewers independently assessed methodological quality using AMSTAR2.
Results Searching identified 361 citations, 45 were potentially relevant, and 7 were included. Reviews were published between 2007 and 2021, and included 47 virus studies. Two earlier reviews (2007, 2016) of 21 studies found sufficient evidence that mechanical ventilation (airflow patterns, ventilation rates) plays a role in airborne transmission; however, both found insufficient evidence to quantify minimum mechanical ventilation requirements. One review (2017) of 9 studies examining humidity and indoor air quality found that influenza virus survival was lowest between 40% and 80% relative humidity; authors noted that ventilation rates were a confounding variable. Two reviews (2021) examined mitigation strategies for coronavirus transmission, finding droplet transmission decreased with increasing temperature and relative humidity. One review (2020) identified 14 studies examining coronavirus transmission in air-conditioning systems, finding HVAC systems played a role in virus spread during previous coronavirus outbreaks. One review (2020) examined virus transmission interventions on public ground transportation, finding ventilation and filtration to be effective.
Discussion Seven reviews synthesizing 47 studies demonstrate a role of HVAC in mitigating airborne virus transmission. Ventilation, humidity, temperature, and filtration can play a role in viability and transmission of viruses, including coronaviruses. Recommendations for minimum standards were not possible due to few studies investigating a given HVAC parameter.
This overview synthesized seven previous reviews that included 47 studies examining HVAC design features and their effects on airborne transmission of viruses, serving as a starting point for future systematic reviews and identifying priorities for primary research.
The strengths of this study include its comprehensiveness and use of methods to avoid bias, such as pre-specification of inclusion/exclusion criteria and involvement of at least two reviewers at all stages.
The main limitation stems from the limits of the included reviews because many reviews did not meet the internationally recognized definitions and methodological expectations of systematic reviews.
Introduction
The 2019 novel coronavirus (SARS-CoV-2) outbreak, first detected in Wuhan, China, was characterized as a pandemic by the World Health Organization (WHO) in March 2020.[1] Fifteen months later (mid-June 2021), the WHO had reported over 177 million cases of the disease (COVID-19) caused by the virus with over 3.8 million deaths worldwide.[2] Throughout the pandemic there have been conflicting views and debate about routes of transmission.[3–6] While an important mode of transmission is close contact (i.e., contact transmission with an infected person or contaminated surface, or droplet transmission resulting from close proximity to an infected person), airborne transmission has not been ruled out. Several recent reviews of the scientific literature have identified evidence indicating airborne transmission, which could be particularly problematic in confined and/or crowded indoor spaces.[7–9] Public health recommendations reflect this emerging evidence with advice to maximize ventilation, ensure proper maintenance and functioning of heating, ventilation and air conditioning (HVAC) systems, and increase use of fresh air where possible.[10]
Airborne transmission occurs as a result of smaller droplets staying aloft longer due to their small size and therefore travelling further due to air currents.[3] Several possible mechanisms of airborne coronavirus transmission exist including: 1) droplet (bioaerosols) generation by infectious persons with coughing, sneezing, breathing and talking, which remain airborne for a period of hours to days; 2) short- to long-range transport through heating, ventilation and air conditioning (HVAC) systems and subsequent inhalation of droplet nuclei by other people; and 3) airborne transport of bioaerosols to surfaces (or 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 (e.g., ventilation rate, air changes per hour (ACH), airflow pattern, pressurization), filtration (e.g., filter minimum efficiency reporting value (MERV) rating, age, extent of use), ultraviolet (UV) radiation (e.g., UV power, 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 occupant health and well-being and to affect timely public health policies. Previous systematic reviews provide a starting point for understanding what is already known from the scientific literature about HVAC systems and airborne transmission of infectious agents. We conducted an overview of reviews to identify and synthesize previous systematic reviews on this topic.
Methods
Standards recommended by the international Cochrane organization for the conduct of an overview of reviews [14] were followed. The research question guiding this work was: what is the current synthesized evidence about the effects of heating, ventilation and air conditioning (HVAC) design features on virus transmission?
Search Strategy
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. Search strategies appear in Supplemental Material 1. The unfiltered search strategies were peer-reviewed by two librarians (TL, 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 which is registered [15] and its protocol is publicly available.[16] Reference lists of 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. References were managed in EndNote; duplicate records were removed prior to screening.
Study Selection
Two reviewers (GMT, LH) independently screened the titles and abstracts of all 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 studies classified as yes or maybe. The same reviewers independently applied inclusion/exclusion criteria to each full text document and classified studies as include or exclude. Discrepancies were resolved through discussion between the two reviewers. Reasons for excluding studies at the full text stage were documented (Supplemental Material 3).
Inclusion and Exclusion Criteria
Inclusion and exclusion criteria are detailed in Supplemental Material 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 transmission of viruses (or similar agents). The HVAC features of interest were: mechanical ventilation (ventilation rate, air change, air exchange, airflow); filtration (air filtration, filter type, minimum efficiency reporting value (MERV) rating, filter age and/or use, pressure drop, holding capacity, replacement, change frequency); ultraviolet germicidal irradiation, UVGI (power, dose, uniformity of dose, flow rate, bioaerosol inactivation efficiency, location); humidity or relative humidity. Inclusion was staged in two ways. First, our primary interest was viruses; however, we searched broadly for any infectious agent. Given that we identified reviews specific to viruses, we excluded those that focused exclusively on other agents (e.g., bacteria). Second, we were initially interested in systematic reviews defined by the international Cochrane organization as reviews that use a pre-defined, systematic approach and follow standard approaches to search the literature, select studies for inclusion, assess methodological quality of included studies, extract and synthesize/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 two or more 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 an HVAC design feature, including ventilation, filtration, ultraviolet radiation, or humidity.
Quality Assessment
Methodological quality of 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, LH) independently assessed the included reviews. Discrepancies were resolved through discussion.
Data Extraction
The following information was extracted from each review: citation information (e.g., authors, year of publication, country of corresponding author); objectives; search strategy; inclusion/exclusion criteria; settings; population characteristics (as applicable); agent studied (e.g., type of virus, bioaerosol, droplet characteristics); HVAC design features studied; number and characteristics of studies relevant to this overview’s research question; results (as reported by review authors) and review authors’ conclusions relevant to this overview’s research question. Our primary outcome was quantitative measures 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 pre-defined form. A second reviewer (EK) verified data. Discrepancies were resolved through discussion and referring to the relevant publication.
Data Analysis
We anticipated that the included reviews would not have conducted meta-analyses. We planned to present results in tabular and narrative form. Tables were created describing the reviews, their results (including any quantitative data of 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 is 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 environment, we only report on studies specific to the indoor (built) environment.
Patient and Public Involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Results
The search retrieved 361 citations, of which 45 were considered potentially relevant and 7 met the inclusion criteria (Figure 1). Tables 1 and 2 provide summaries of the included reviews. The reviews varied somewhat in their objectives (e.g., investigate mechanical ventilation, ventilation rates, airflow patterns, effects of humidity, stability of aerosols containing coronaviruses), agents (e.g., coronaviruses, influenza, viruses), and settings (e.g., built environment, healthcare settings, public ground transportation). 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, Supplemental Material 4).
Flow of studies through the selection process
Li et al [13] examined the role of ventilation (specifically ventilation rates and airflow patterns) in airborne transmission of infectious agents in indoor settings. The authors included 40 English language studies overall, with 16 specific to viruses reported between 1962 and 2005 (median year 1985/1996). Three of the 16 studies included multiple papers (Table 3) which increased the total count to 21. Sixteen studies were epidemiologic, four involved other observational designs, and one was experimental. Three studies had limited and four had no investigation of ventilation rates or airflow. Studies involved a variety of settings: hospital, hospital ward, or health clinic (n=9); aircraft (n=3); nursing home (n=3); school (n=2); high-rise apartments (n=2); offices (n=1); and animal cage (n=1). Viral agents included SARS-CoV-1 (n=7), influenza (n=5), measles (n=4), chicken pox (n=2), rhinovirus (n=1), common cold (n=1), and smallpox (n=1). Overall quality was assessed as good for twelve, average for five, and unsatisfactory for four studies. The researchers convened a panel of experts in medicine, public health, and engineering. They used a modified Delphi approach with final consensus meeting to rate the ‘evidentiary threshold’ to support their hypothesis, i.e., direct contribution of ventilation to airborne transmission. Among the virus studies, eight were rated as conclusive, while eight were partly conclusive, and five were non-conclusive. Among the eight conclusive studies, two examined ventilation rates showing higher rates of infection for influenza with lower ventilation rates and six demonstrated an association between airflow patterns and transmission of measles (pediatric office suite), chicken pox (hospital), smallpox (hospital), and SARS-CoV-1 (hospital). In all studies the aerosols travelled a “considerable distance” which the reviewers note “seemed to be related to building design” [13, p.12] (e.g., placement of heating radiator, room pressure, functional status of return air outlet). None of the virus studies provided data to support “specification and quantification of the minimum ventilation requirements”.[13, p.3]
Luongo et al [19] examined evidence from epidemiological studies for the association of mechanical ventilation (at least one HVAC parameter) with airborne transmission of infectious agents in buildings. While the authors included 13 English-language studies, 3 were specific to viruses; all studies were observational and were reported between 1996 and 2011 (median year 2004). One study had two papers which increased the total count to 4 (Table 3). All four virus studies were also included Li et al.[13] Settings included nursing home (n=2), office buildings (n=1), and hospitals (n=1). Viruses represented in the studies included influenza (n=2), SARS-CoV-1 (n=1), and rhinovirus (n=1). Review authors did not assess methodological quality but provided a narrative commentary of the strengths and limitations of each study. Two of the four studies found an association between virus incidence rates, self-reported incidence rates, and risk of exposure with HVAC design features. In a retrospective cohort study of a SARS-CoV-1 outbreak in a hospital, authors measured ventilation rates and found that “proximity to index patient associated with transmission”.[19, p.668]. Authors of the second study blindly adjusted outdoor air supply dampers in three office buildings and found a significant positive association between average CO2 greater than 100 ppm above background and frequency of rhinovirus detection in air filters. The third study found lower incidence of influenza in newer nursing homes that had 100% outside air delivery (compared to older homes with 30-70% recirculated air) and filtered room supply (compared with no filtration) during one season; however, data collected over five subsequent years, reported in the fourth study, found no clear association. None of the studies quantified minimum ventilation requirements.
Derby et al [20] conducted a literature review to assess the effects of low humidity (less than or equal to 40% relative humidity or RH) on comfort, health, and indoor environmental quality. While the review included around 70 papers, 9 papers examined the effects of humidity on the viability or transmission of airborne viruses. Seven studies were experimental (involving laboratory testing), one paper was a re-analysis of data from one of the experimental studies, and one study involved modelling. Most studies focused on influenza, with one 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 relative humidity from approximately 5-25% RH at the lower range to 75-100% RH at the upper range. The absolute humidity was approximately 25 g/m3 or less in all studies except one (which ranged from 25 to 125 g/m3). Review authors did not assess the methodological quality of the included studies. In terms of virus viability, four studies showed a reduction at mid-range RH (i.e., approximately 50% RH). Review authors further noted that five studies showed that “virus survival exhibited a canonical dip between 40 and 80% RH”,[20, p.39] and that in almost all cases decline in survival was correlated with increased length of exposure. Three studies examined influenza transmission. One study showed reduced transmission among guinea pigs at 50% 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 five times more infectious virus at 7-23% RH versus greater than 43% RH. A modelling study of influenza virus transmission via coughing showed that the infectious virus concentration was 2.4 times more at 10% RH versus 90% RH after 10 minutes and the ratio increased over time. They also demonstrated that the effect of humidity related to droplet size: settling of larger droplets and inactivation of smaller droplets (<5 µm) with greater humidity. They concluded that inactivation resulting from high relative humidity coupled with ventilation is important to remove smaller droplets.
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, p.2] The authors identified 14 studies published between 2003 and 2020 (11 peer-reviewed, 3 pre-prints all concerning SARS-CoV-2); studies investigated outbreaks in Hong Kong (n=7), South Korea (n=1), Japan (n=3), and China (n=3). Seven studies examined two outbreaks associated with SARS-CoV-1: five studies examined outbreaks (different areas or groups of individuals) within the same hospital and 2 studies investigated an outbreak in the same private high-rise housing estate. Five of the SARS-CoV-1 studies from Chirico et al [21] are shared references with Li et al [13] (Table 3). Review authors indicated that six of the seven studies indirectly demonstrated a role of the HVAC system (through epidemiological data, spatiotemporal patterns of infection, or modelling). One study investigated an outbreak of MERS-CoV in a hospital setting and demonstrated contamination of the HVAC system through environmental sampling.[22] Six studies investigated outbreaks of SARS-CoV-2: one study examined 318 outbreaks in 120 cities in China including community and workplace settings; three studies examined an outbreak on a ship in Japan; and two studies examined the same outbreak in a restaurant. Three observational studies suspected a role of the HVAC system, two studies (both of ship outbreak) did not find evidence of a role of HVAC based on the spatiotemporal distribution of cases, and one study (of restaurant outbreak) showed support of the role of HVAC by computer simulation. Review authors indicated that they were not able to appropriately evaluate the quality of the included studies. Review authors concluded that there is sufficient evidence from SARS-CoV-1 and MERS-CoV demonstrating a role of HVAC in 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. While there is a lack of evidence for SARS-CoV-2, there is not 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, p.478] The authors searched for studies published since 2000 and identified four relevant studies, published between 2007 and 2016, including a systematic review, a case-control study, and two modelling studies. The systematic review by Browne et al [24] identified 41 studies examining the risk of transmission of Influenza A (H1N1/09) (29 studies), SARS-CoV (5 studies), or MERS-CoV (2 studies) related to sea (6 studies), ground (6 studies), or air (30 studies) transport. Zhen et al [23] summarized results from four quantitative studies included in Browne et al,[24] and concluded that “use of public transport increased the risk of influenza transmission”.[23, p.481] Zhen et al [23] identified a multi-centre 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 use of train, aeroplane or taxi. The case-control study was assessed as moderate risk of bias by Zhen et al [23]; risk of bias was not reported for the other three studies. Zhen et al [23] also identified two modelling studies, one estimated the reproduction number for influenza infection in a train and the other tested simulations to predict influenza infection probability for four bus ventilation systems. The first modelling study showed that masks could decrease reproduction number resulting in lower risk of disease transmission, with HEPA more effective than surgical masks. Further, doubling the ventilation rate reduced risk similar to use of HEPA masks, and was considered more feasible and cost-effective. The second modelling 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 non-recirculated air cannot be provided. Given the limited number of research studies, Zhen et al [23] also identified and discussed national and international guidance documents, e.g., World Health Organization, etc.[25–31] While general recommendations have been made to reduce risk (e.g., minimize use of public transport, environmental controls, respiratory etiquette, hand hygiene, mask use), there is no indication of the empirical evidence specific to these measures, in particular mechanical ventilation.
Da Silva et al [32] 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”.[32, p.2] The review authors identified 11 studies: eight studies examined the viability of coronaviruses in air samples but review authors did not describe the relationship with HVAC features, including one MERS-CoV study [22] which was described above by Chirico et al.[21] Three 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.[33] The other MERS-CoV study examined virus inactivation under two conditions [34]: 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.”.[32, p.5-6]. One study showed aerosol survival time for SARS-CoV-2 of 3 h.[35] Review authors did not assess the methodological quality of 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”.[32, p.8] However, their conclusions are based on studies of both indoor and outdoor environments.
Noorimotlagh et al [36] performed a systematic review of SARS-CoV-2 literature “to collect all available studies concerning inactivation methods, environmental survival, and control and prevention strategies”.[36, p.1] While 42 studies were identified, four provided information on temperature and humidity in the built environment, investigating, MERS-CoV (n=2), SARS-CoV-1 and SARS-CoV-2 (n=1), and Phi6 (n=1) which is a bacteriophage used as a surrogate for virus. All were laboratory-based experimental studies. Review authors did not assess or comment on the methodological quality of the studies they included. The aerosolized MERS-CoV data from van Doremalen et al [33] was reported in da Silva et al [32] although not extracted by Noorimotlagh et al.[36] A MERS-CoV study found robustness and strong capability to survive at 25°C and 79% RH;[34] this was a shared reference with da Silva et al.[32] While da Silva et al [32] reported the SARS-CoV-2 survival time from van Doremalen et al [35], Noorimotlagh et al [36] further clarified that the aerosol survival time for SARS-CoV-2 of 3 h was at 40% RH and 21-23°C, and that the stability of SARS-CoV-2 was similar to SARS-CoV-1.[35] The Phi6 study showed highest virus survival at greater than 85% RH and less than 60% RH with significant decrease between 60% RH and 85% RH.[37] At fixed humidity of 75% RH, infectivity decreased two orders of magnitude between 19°C and 25°C.[37] The review authors concluded that “temperature and relative humidity are important factors in the survival of SARS-CoV-2” [36, p.1] and that “disease transmission via droplets is inhibited by increasing both temperature and RH in buildings”.[36, p.11] One review recommendation was “proper ventilation of the buildings in time of aerosol generating”;[36, p.13] however, studies of ventilation were not reviewed.
The network of the seven included reviews and their 47 references relevant to this overview are shown in Figure 2. Derby et al [20] and Zhen et al [23] share no references with the five other reviews. da Silva et al [32] and Noorimotlagh et al [36] and share three references regarding experimental studies of MERS-CoV and SARS-CoV-2 [33–35] (Table 3). Also, da Silva et al [31] shares one 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], share a reference on SARS-CoV-1 in hospital wards [38]. Li et al [13] and Chirico et al [21] share other related references on SARS-CoV-1 in hospital wards [39, 40]. Similarly, Li et al [13] and Luongo et al [19] share studies on influenza in nursing homes [41, 42] and rhinovirus in offices [43]. Li et al [13] and Chirico et al [21] share two other references regarding SARS-CoV-1 in high-rise apartment complexes [44, 45]. Overall 12 references were shared between the seven included reviews and eight of the shared references were shared with one review, Li et al [13] (Figure 2).
Network representing relevant references (grey circles) from the seven included reviews (black circles): 1_Li_2007; 2_Luongo_2016; 3_Derby_2017; 4_Chirico_2020; 5_Zhen_2020, 6_daSilva_2021; 7_Noorimotlagh_2021.
Shared references:
1_Li_2007, 2_Luongo_2016 and 4_Chirico_2020 share Wong_2004 1_Li_2007 and 2_Luongo_2016 share Drinka_1996; Drinka 2004; Myatt 2004
1_Li_2007 and 4_Chirico_2020 share Li_2005_a; Li_2005_b; Yu_2004; Yu_2005 4_Chirico_2020 and 6_daSilva_2021 share Kim_2016
6_daSilva_2021 and 7_Noorimotlagh_2021 share vanDoremalen_2013; Pyankov_2018; vanDoremalen_2020
Table 4 provides the assessments of the methodological quality of the reviews based on AMSTAR2. Three papers described themselves as systematic reviews, two were rapid reviews, one was a broad literature survey, and one 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 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 sources of funding for included studies. None of the reviews conducted a meta-analysis; all provided a narrative synthesis of results/observations across included studies. A previous review [19] spoke about the need for more well-designed studies (including representative sampling, clear and consistent measurement methods and reporting of data) with the goal of meta-analysis to integrate results.
Discussion
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 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 one study showing tuberculin conversion was significantly associated with ventilation rates of less than 2 ACH in general patient rooms.[44] Published in 2007 shortly after the 2003 SARS-CoV-1 epidemic, Li et al called for a “multidisciplinary research culture” [13, p.14] to study outbreaks, as well as smaller scale transmission occurrences, to fill the gap with respect to quantifying minimum ventilation standards in both clinical and non-clinical settings. A subsequent review by Luongo et al [19] published almost ten years later in 2016 included a subset of four of the virus studies identified by Li et al [13] with similar conclusions about the possible association between ventilation features (low outdoor air supply, 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 multi-disciplinary epidemiological studies. More recently in response to the current COVID-19 pandemic, Chirico et al [21] examined HVAC systems and their role in airborne transmission of coronaviruses; they concluded there was sufficient evidence demonstrating an association for SARS-CoV-1 and MERS-CoV while 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 virus survival was lowest between 40% and 80% RH and survival time decreased with length of exposure to humidity. One of the studies from Noorimotlagh et al [36] indicated that aerosolized SARS-CoV-2 can survive for 3 h at 40% RH and 21-23 °C [35]. In another recent review published in 2021, da Silva et al [32] examined mitigation strategies and found two studies demonstrating that droplet transmission decreased with increasing both temperature and relative humidity in buildings. A recent review (2020) by Zhen et al [23] examined interventions to reduce virus transmission on public ground transportation; two modelling studies showed ventilation and filtration to be effective.
While there is an extensive body of literature examining HVAC and its role in airborne virus transmission, there is lack of empirical evidence to quantify minimum standards for HVAC design features in the built environment. The previous reviews have discussed this gap, stressed the need for methodologically rigorous epidemiological studies involving multiple disciplines (e.g., engineering, medicine, epidemiology, and public health), and discussed considerations for future research. These include specificity of the virus, its construction and envelope composition, the infectious dose, and the droplet size containing the virus. Review authors have called for standardizing experimental conditions, measurements, terminology, and reporting, as well as simulating real world conditions.[19,20,32] 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, p.33] They further noted that the number and complexity of variables to consider “greatly increases the test matrices required” [20, p.39] 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 (e.g., type and intensity of respiratory activity) of occupants, time spent within a space, occupancy and occupant density. Despite specification of air flow parameters, the flow of air in occupied spaces is almost always turbulent (versus laminar) such that particles “are constantly mixing and moving in varied ways across a space” making assessments and predictions challenging.[47] Finally, research results need to be interpreted in light of technological differences in HVAC systems around the world.[21] Engineers have developed sophisticated methods (through modelling, computational fluid dynamics, etc.) that allow for 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. As well, these studies may isolate one component in the chain of transmission which does not necessarily equate to actual disease (e.g., detection of viral particles vs infectivity vs disease outcomes).[19,20,32] Results from modelling studies need to be considered alongside epidemiological studies. Previous reviews have highlighted many challenges with studying outbreaks: Li et al [13] mentioned 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, p.14] They proposed as a solution “contemporaneous air-sampling and environmental measurements” [13, p.14] 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 use of methods to avoid bias, such as pre-specification of inclusion/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. While 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 their included studies which is key to determining the validity and certainty of available evidence. We also did not find reviews of all HVAC design features, for example, none of the included reviews examined UVGI (though a recent narrative review has been published in the context of COVID-19 [48]) and only a small number of studies across the reviews examined filtration.
The findings of this overview provide several implications for public health measures to mitigate spread of viral transmission in buildings. First, ventilation rates and air flow patterns have been shown to associate with virus transmission. Second, humidity and temperature associate 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] Reviews 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 inter-disciplinary research with a specific focus on quantifying minimum specifications for HVAC features. While one review did not find sufficient evidence of HVAC and airborne transmission specific to SARS-CoV-2, authors did advise (based on evidence for MERS-CoV and SARS-CoV-1) that attention be given to design and management of HVAC systems as a precautionary measure until further evidence indicates otherwise.[21]
Conclusion
There is growing evidence of airborne transmission of different viruses including coronaviruses and specifically SARS-CoV-2 which has been the source of immense global impacts in terms of morbidity, mortality, and peripheral effects of pandemic restrictions. HVAC systems and their specific features have the potential to mitigate transmission in the built environment: there is evidence that ventilation rates, airflow patterns, humidity, temperature, and filtration can influence virus transmission. Enhancing HVAC systems in the built environment (including schools, office buildings, commercial spaces, recreation centers, transport vehicles, etc.) 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 behaviour and result in other (e.g., social) consequences will be more sustainable.[21]
Data Availability
No additional data available. All data are available in the manuscript and the supplemental material.
Contributors
GMT contributed to design, screening and study selection, data extraction and synthesis, and drafting the manuscript. EK contributed to data extraction and synthesis, and drafting the manuscript. BAF contributed to design, interpretation of results, and critically reviewed the manuscript. LZ contributed to design, interpretation of results, and critically reviewed the manuscript. LH conceived of the study, contributed to design, screening and study selection, data extraction and synthesis, and drafting the manuscript.
Funding
This work is funded by a Canadian Institutes of Health Research (CIHR) Operating Grant: Canadian 2019 Novel Coronavirus (COVID-19) Rapid Research Funding Opportunity [https://webapps.cihr-irsc.gc.ca/decisions/p/project_details.html?applId=422567&lang=en; grant number: not applicable]. Dr. Hartling is supported by a Canada Research Chair (Tier 1) in Knowledge Synthesis and Translation. The funders had no role in the project.
Competing Interests
Dr. Fleck is an unpaid advisor for Pura Air Inc. in Vancouver. None of the other authors has competing interests.
Data Sharing
No additional data available. All data are available in the manuscript and the supplemental material.
Acknowledgements
We thank Tara Landry and Alison Henry for conducting the peer review of the unfiltered search strategies. We thank Liz Dennett for providing the filter for systematic reviews for Ovid MEDLINE.