COVID-19 Outbreak and Hospital Air Quality: A Systematic Review of Evidence on Air Filtration and Recirculation.

The outbreak of SARS-CoV-2 has made us all think critically about hospital indoor air quality and the approaches to remove, dilute, and disinfect pathogenic organisms from the hospital environment. While specific aspects of the coronavirus infectivity, spread, and routes of transmission are still under rigorous investigation, it seems that a recollection of knowledge from the literature can provide useful lessons to cope with this new situation. As a result, a systematic literature review was conducted on the safety of air filtration and air recirculation in healthcare premises. This review targeted a wide range of evidence from codes and regulations, to peer-reviewed publications, and best practice standards. The literature search resulted in 394 publications, of which 109 documents were included in the final review. Overall, even though solid evidence to support current practice is very scarce, proper filtration remains one important approach to maintain the cleanliness of indoor air in hospitals. Given the rather large physical footprint of the filtration system, a range of short-term and long-term solutions from the literature are collected. Nonetheless, there is a need for a rigorous and feasible line of research in the area of air filtration and recirculation in healthcare facilities. Such efforts can enhance the performance of healthcare facilities under normal conditions or during a pandemic. Past innovations can be adopted for the new outbreak at low-to-minimal cost.

Introduction

The prevention of healthcare associated infections (HAIs) has long been a top strategic priority for the Center for Disease Control and Prevention (CDC).[1] Still, an estimated 90,000 deaths occur per year in the United States.[2] The recent pandemic manifested an urgent need to better understand and implement design, maintenance, and operations that ensure indoor air quality in healthcare facilities. Specifically for the SARS-CoV-2 virus, the CDC recognizes three main routes of transmission: (1) direct large droplet transmission between people within close proximity; (2) indirect respiratory droplet deposition on surfaces and objects and subsequent transmission via the contaminated fomite; and (3) the airborne transmission via small particle aerosols containing viable virus. Although the portion each route contributes to disease transmission is under close investigation, one can reasonably assume that lowering the number of virus particles in the indoor space would result in lower rates of transmission. Engineering tools can be instrumental to remove, contain, and dilute virus concentrations in the spaces where the COVID-19 patients are evaluated and treated, and air filtration–recirculation is a long-lasting method used to remove and dilute contaminants.[3−5]

What Do Past Outbreaks Reveal?

Infection may transfer via air from one person to another and become epidemic.[6] Although it is accepted that the airborne route exists, its attributed contribution to the spread of infectious diseases is less certain.[7] For example, tuberculosis (TB) is transmitted person-to-person by the airborne route via small particle aerosols measuring 1 to 5 μm, sufficiently small that they remain aloft in the air.[8] In 1966, aboard the Navy ship U.S.S. Richard E. Byrd, 45.5% of the at-risk crew members (140 out of 308) unfortunately became infected with TB. In this episode, air recirculation on the ship was implicated as the primary amplification pathway of infection.[9] In 1972, Ehrenkranz reported 21 healthcare staff who developed latent TB infections out of 60 formerly tuberculin-negative personnel. Ten of these 21 staff had little or no direct contact with the patient with TB, and the authors concluded that they likely had been infected by the spread of Mycobacterium tuberculosis through an unbalanced air conditioning system which was found by smoke stick to allow air from patients’ room to enter the hallway, and this system lacked a “high-efficiency filter” (80% or greater).[7,10] During 1983–1984, 6 employees of the emergency department of Parkland Memorial Hospital (Dallas, TX) developed active TB. One of the reasons that led to the outbreak of TB was the lack of adequate filtration of the 60% recirculated air which could not remove tubercle bacilli in the emergency department.[11] Prior to 1994, it was recommended that isolation rooms for patients known or suspected of active TB disease should have had a minimum total of 6 air changes per hour (ACH), at least 2 of which with outdoor air.[12−14] In the updated 1994 guidelines, the CDC recommended that, wherever feasible, air flow rates be increased to 12 air changes per hour with recirculating air in the room through fixed, monitored high efficiency particulate air (HEPA) filtration or portable air cleaners.[15]

Reviews of other environmental parameters in healthcare facilities exist, such as Memarzadeh (2012) on the effect of temperature and relative humidity on pathogen viability.[16] Another review specifically targeted thermal comfort in hospitals, concluding that while it is almost impossible to satisfy every occupant, there are certain engineering tools and numerical techniques that can be utilized to improve the design and operation of the hospital.[17] Most recently, a review of scientific evidence on hospital buildings studied the temperature, relative humidity, and ventilation system. It briefly mentioned the filtration systems; however, recirculation was not considered.[18]

Focus of the Current Paper

The current state of healthcare ventilation management stems from two complementary schools of thought around minimizing both recirculation of indoor air and energy demands. Various studies demonstrate the importance of the airborne route of infection transmission in the hospital setting.[19] On one hand, it seems unreasonable to allow recirculation of return air into patient care environments. On the other hand, it seems reasonable to take advantage of the energy in the return air to dilute contamination through air recirculation, while ensuring that the return air is clean and comfortable. With recent advancements in technology, the latter can also be achieved via energy recovery techniques, and recirculation is no longer the sole option. However, these techniques are not always a feasible and affordable retrofit response to the emergence of COVID-19, especially in existing facilities while replacing the filtration system, and altering the air distribution system could be more convenient. Perhaps, a natural question is, to what extent can new approaches in air filtration/recirculation improve the health and safety of patients and the front-line medical personnel? These approaches bring about this literature review on the issue of air filtration and recirculation in healthcare facilities. In the present work, we will first briefly discuss the theoretical background of interconnections between air filtration and recirculation. Next, the process of identifying relevant publications, followed by assessing their methodology, will be explained, and evidence for air filtration and air recirculation and potential interconnections will be reviewed.

Theoretical Background

The true effect of air filtration and air recirculation on the temporal and spatial distributions of contaminants indoors is a very difficult problem—one that does not have an analytical solution and is computationally very expensive. However, one could set aside the spatial term by assuming that contamination is uniformly distributed throughout the space. This so-called well-mixed assumption is a useful simplification that is often used in the field.[20,21] Taking the room as a large control volume, one could derive the differential equationwhere V is the volume of the room, C is contaminant concentration as a function of time, Cs is contaminant concentration at the supply air inlet, Qs and Qr are supply and return flow rates, and S is a source of contamination in the room (Figure ). Without a loss of generality, let us assume that the supply and return flow rates are identical, Qs = Qr = Q. It is noteworthy that for pressurized spaces (i.e., operating rooms (OR), AIIRs) this assumption is not valid. In general, eq is a linear differential equation with an exponential solution. For example, eq shows an exponential decay for a single release of contaminant. One can argue that the efficiency of the filter changes with time, which will turn eq into a nonlinear problem that must be solved numerically.[22]

Figure 1

Contaminant distribution in the space under the well-mixed condition.

Moreover, the concentration at the supply air inlet can be written aswhere α is the percentage of outside air mixed with return air, Cout is contaminant concentration in the outside environment, and η is filter efficiency. For the scope of this work, Cout can be assumed as zero, as one does not expect to have COVID concentrations at the outside air intake location. In essence, α represents the recirculation in the system where α = 100% shows a full recirculation system and α = 0% shows a one-pass (i.e., no recirculation) system. Combining eq and eq , and moving V to the right-hand side of the equation, we will havewhere the term is the space ventilation rate that is often referred to as air changes per hour (ACH). To show the effect of filtration against increasing the ratio of outside air, eq was solved for different values of α (10% to 100%) and η ranging from MERV 12 (70% efficiency for 0.3 μm) to HEPA (99.97% efficiency for 0.3 μm). To compare different cases, a time index (T) was defined as the amount of time it takes for the concentration of the contaminant to decay to 0.001 initial release value. The shorter the time index, the more efficient the ventilation system is in removing contamination from the space. Results showed that for low filter efficiencies T reduced significantly by introducing more outside air (α on the x-axis), but HEPA filters can clean the space such that T became insensitive to the portion of outside air (Figure ).

Figure 2

Time index (T) for various outside air ratios and filter efficiencies (ACH = 12).

Another interesting interaction between the filtration and recirculation systems is that filters with higher efficiencies create a pressure drop across the filter, leading to lower supply airflow rates.[23,24] This process is more profound for fibrous media that are frequently used for HVAC air filtration.[25,26] The literature suggests that the pressure drop across the air filter is linearly related to filter efficiency.[27,28] Further, for a constant fan power, pressure drop is inversely related to flow rate squared.[29] Therefore, for a room with a fixed volume and fan power:The combination of eq and eq was solved to obtain the coupled effect of air filtration and recirculation. As shown in Figure , considering the impact of higher filter efficiency on the recirculation flow rate can result in significant increase in the cleaning time index (T). It must be noted that there are several other filtration mechanisms that produce less significant pressure drops compared to fibrous filter media,[30] yet the coupled effect seems to deserve great consideration when increasing filter efficiency without changing fan power.

Figure 3

Effect of pressure drop on cleaning time–filtration and recirculation coupling (α = 10%).

Method

In Phase I of this systematic literature review, related academic databases, and codes and guidelines by relevant organizations have been identified. The databases include PubMed, Web of Science, Engineering Village, and Compendex. Key terms such as “filtration”, “recirculation”, “ventilation”, “HVAC”, “mechanical system” in combination with terms such as “hospital”, “healthcare facilities”, “outbreak”, “epidemic”, and “pandemic” were used to identify relevant articles from the above databases. Table shows the results of the initial search of the key terms with a separate focus on air filtration and air recirculation in healthcare facilities. Next, the research team looked for duplicated outcomes from these databases. The reference list of these articles were also searched to include relevant articles that otherwise may not have been included. The first round of source selection resulted in 394 publications.

Table 1

Result of Initial Search of the Keyword Combinations for Air Filtration and Recirculation

database	air filtration	air recirculation
PubMed	403	36
Web of Science	132	11
Engineering Village	62	45
Compendex	111	23

Only peer-reviewed documents, available in English, were selected for further review. Next, these publications were independently reviewed by two experts for relevance to the topic of this literature review, that is, the effect of air filtration and air recirculation on the spread of pathogenic agents in the hospital. The inclusion criteria were as follows: availability in English (390 documents), peer reviewed conference or journal articles (307 documents), a focus on the building mechanical ventilation (178 documents). Finally, only those publications with either original results or rigorous review papers were included for the final review (91 documents). Figure illustrates the systematic review method.

Figure 4

Flow diagram of record identification, eligibility, and inclusio. Chart style courtesy of PRISMA.[31]

In Phase II of document identification, guidelines from relevant organizations such as the Facility Guide Institute (FGI), The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE), The American Hospital Association (AHA), the American Society of Healthcare Engineers (ASHE), and the Center for Disease Control (CDC) were included. Guidelines from regulatory bodies from across the globe were also included, contingent upon the availability of the documents in English. This study reviews 20 codes, guideline, standards, and handbooks pertaining to ventilation in healthcare facilities.

Further, the level of evidence was evaluated based on the methodology in the articles (Table ). To that end, the publications were categorized in five levels based on the methodology rigor based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations. Of note, none of the studies retrieved were randomized trials with direct evidence. Also, Figure shows the distribution of publications with original data based on the methodology employed.

Table 2

Level of evidence assessment matrix

level of evidence	count	criteria
Level 1	0	clinical trial randomized studies with direct evidence
Level 2	27	physical or biological field experiment repeated with/without computational fluid dynamics (CFD)
Level 3	31	Scale model/chamber experiment with or without CFD
Level 4	10	CFD and other simulation modeling
Level 5	23	Literature review/expert opinion

Figure 5

Methodology of publications with original data.

Codes, Guidelines, and Standards

Air Filtration

As it relates to air filtration, ASHRAE recognizes five different spaces in a typical in-patient hospital facility: (1) surgery and critical care spaces including operating rooms (ORs), inpatient and ambulatory diagnostic and therapeutic radiology, and inpatient delivery rooms; (2) inpatient nursing including airborne infection isolation rooms (AIIRs), intermediate, critical, and intensive care units; (3) protective environment rooms, or rooms that are intended to protect a high-risk immunocompromised patient from human and environmental airborne pathogens; (4) laboratories and procedure rooms and associated semirestricted spaces; and (5) all other service, administrative, food preparation, and laundry, and storage spaces.[32] It must be noted that in the most recent version (2017), ASHRAE Standard 170 offers design parameters for in-patient facilities, out-patient facilities, and nursing homes. However, the above categories and the associated requirements are still very similar for these three facility types. Specifically, filtration and air recirculation requirements for inpatient facilities, where a COVID patient is likely hospitalized, have not changed compared to the previous version. Moreover, in the older versions of Standard 170, ORs are classified based on the type of surgery in Group A (minor surgery), Group B (procedures that require intravenous sedation), and Group C (procedures that require general/regional block anesthesia). This classification has been removed from the 2017 revision, and the definition by the Facilities Guideline Institute (FGI) has been adopted.[33] In this paper, however, we still note the old classification just for referencing purposes. Table shows the minimum filtration and air recirculation requirements for these spaces based on the various standards reviewed in this work. The filtration rating is reported in the minimum efficiency reporting value (MERV) rating, that is, the lowest point of filter efficiency in removing three different size ranges of 0.3–1.0 μm, 1.0–3.0 μm, and 3.0–10.0 μm.[34] MERV-equivalent efficiencies are reported for the non-US standards.

Table 3

Code and Standard Requirements for Filtration and Air-Recirculation in Patient Care Areas

space	standard	filtration	air recirculationa
Surgery and critical spaces	United States	MERV 14	Y
	United Kingdom	MERV 13	NR
	Germany	MERV 15–16	NR
	Spain	MERV 15–16	Y
	Canada	HEPA	Opt.
Inpatient spaces (AIIR, ICU, CCU)	United States	MERV 14	N
	United Kingdom	MERV 13	N
	Germany	HEPA	N
	Spain	MERV 14	N
	Canada	MERV 13	Y
Protective environments	United States	HEPA	Y
	United Kingdom	Not covered	NR
	Germany	HEPA	Y
	Spain	MERV 16–17	NR
	Canada	HEPA	Y
Laboratories and procedure rooms	United States	MERV 13	Y
	United Kingdom	MERV 13	NR
	Germany	MERV 15–16	Y
	Spain	MERV 15–16	NR
	Canada	MERV 13-HEPA	Y
Outpatient rooms	United States	MERV 7	Y
	United Kingdom	MERV 6–8	Y
	Germany	MERV9–10	Y
	Spain	MERV 6–8	Y
	Canada	MERV 7	Y

Y:yes; N:no; NR: not required; Opt.: optional.

Ninomura et al. (2001) provided the first edition of the “Guidelines for Design and Construction of Hospital and Health Care Facilities” that the Facility Guidelines Institute (FGI) supported.[35] These guidelines required the protective environment rooms (i.e., dedicated to immunocompromised patients) to be equipped with HEPA filters with a 99.97% efficiency for particles larger than 0.3 μm.[35] For product protection in pharmacies, laminar flow with a nonhydroscopic filter at 99.97% removal efficiency should be designed. The committee on operating room environment of the American College of Surgeons advised the use of HEPA filters as standards for all operating rooms, to help prevent surgical site infections.[36] When two banks of filters are necessary as with areas for inpatient care, and protective environment rooms, the first filter should be on the upstream of air conditioning equipment and the second filter should be downstream of fans or blowers. Hoods that are used to process radioactive or infectious materials should be equipped with HEPA filters. When designing a filtration system, it is important to make sure that no stagnation or short-circuiting of air flow happens.[35]

Patients on airborne precautions for possible aerosolizing infections shall be placed in hospital rooms with HEPA filters to reduce their risk of cross-contamination.[37,38] The American Institute of Architects (AIA) and the CDC recommend that for Airborne Infection Isolation Rooms (AIIRs), the exhaust air should be directly discharged to the outside or filtered prior to recirculation using in-duct HEPA filters.[39,40] The exhaust should be at least 25 feet away from public areas. In case a good location for exhaust cannot be met, HEPA filters can be installed in the discharge duct upstream of the exhaust fan to clean the air.[41] It must be noted that adding a HEPA filter to an existing system is not recommended as it can cause friction in the airstream that leads to delivering less-than-designed amounts of air to the room (i.e., pressure drops).[41] Therefore, if the installation of in-duct HEPA filters is not feasible, a recirculating HEPA filter unit should be added. They are available as wall, floor, and ceiling mounted types. The wall and ceiling units are less obstructive to the airflow, while the floor units are easier to service.[40]

The Australian guideline for design of isolation room by the Victorian Advisory Committee on Infection Control requires the supply or exhaust air to be filtered at their maximum pressure drop.[42] For class P or positive pressure room, the air change rate should be kept at 12 ACH with supply air HEPA filtration.[43] The state of California requires that manufacturers specify filter efficiency and that specified rooms have filters with specific filtration efficiency based on ASHRAE Standards.[44]Table shows filtration and air recirculation requirements for several key spaces in the hospital informed by English et al.[45] and Betz et al.[46] Filter efficiencies for other countries are converted into the MERV rating system.[45−47]

Air Recirculation

Guidelines for Design and Construction of Hospital and Health Care Facilities note there should be no recirculation of air within soiled workrooms or soiled holding areas, ETO-sterilizer rooms, soiled or decontamination rooms, clean workrooms, food preparation centers, ware-washing, soiled linen (sorting and storage) rooms, soiled linen, and trash chute rooms and janitor’s closets.[35] No recirculation of air is allowed in operating rooms, delivery rooms, recovery rooms, or rooms with odor problems.[48] CDC recommends placing patients who have infections which may be transmitted by small particle aerosols in rooms without air recirculation.[49] The state of Texas recommends not allowing recirculation in operating rooms and delivery rooms. If recirculation is necessary to increase the air changes per hour, filters of MERV 17 or higher (HEPA filters) shall be used.[50] These recommendations result in ethical challenges to performing prospective controlled trials. Hence, there is a dearth of high quality studies that prospectively examine the effect of recirculation on disease spread in the hospital setting.[37,51] Before recent pharmacological advancements, anesthetic gases were at risk of combustion, and recirculation of air was prohibited by both federal and state regulations.[45]

A very important note pertaining to the COVID-19 outbreak is that many healthcare facilities, especially those in more populated states (e.g., New York, California), have rapidly surpassed their capacity for isolation rooms. This means that new admissions must be placed in intensive care units (ICUs) or general patient rooms where far less stringent air filtration schemes are required by codes. Awareness must be raised that such practices may exacerbate the spread of virus in the room and other adjoining spaces (i.e., hallways), and the potential to reconfigure air and pressure flows to increase safety capacity should be evaluated.[52]

Epidemic Task Force Guidance

Since the beginning of the COVID-19 pandemic, several organizations such as the World Health Organization, U.S. Federal Government, and State Jurisdictions have established taskforces to assess various aspect of the pandemic such as emergency medical services, health inequalities, and healthcare preparedness. The one taskforce related to hospital ventilation design has been established by ASHRAE. Among other considerations, ASHRAE recommends increasing filtration level (if possible), minimizing air recirculation, maintaining 40–60% relative humidity, and utilizing portable HEPA filtration and anteroom units. Further, ASHRAE proposes five options of airflow alteration to the patient room before hosting a COVID case (Figure ). It must be noted that filtration and recirculation are central to these air distribution schemes. In the case of no filtration (i.e., option 2), air must be discharged at least 25 feet away from any building opening and entrance.

Figure 6

Expedient patient isolation room configurations as suggested by ASHRAE Epidemic Taskforce.

Filtration in Healthcare Facilities

Several research groups examining vulnerable patient populations showed that private rooms with air filtration significantly reduced the risk of infection and mortality.[53−55] For instance, Oren and colleagues (2001) reported that leukemia patients had an infection rate of 50% to invasive pulmonary Aspergillosis in rooms with natural ventilation during hospital renovation. Some patients were then moved to rooms with HEPA filtration and showed no sign of infection even after 3 years.[56] Annual tests performed on all the HEPA filters in a biological laboratory for 13 years showed that they were very reliable. Little failure (∼10%) was observed through the filter medium, provided filter clogging did not occur. Clogging was associated with 75% of failure cases.[57] HEPA filters have a 99.97% or larger removing efficiency of particulate matters of 0.3 μm or larger and are typically protected with less expensive filters on the outside air intake and air return.[58] They were first developed for war time, for their capability to reduce air contamination, along with recirculation of air and laminar flow.[59] It is worth mentioning that not every study supports the use of HEPA filters. Kowalski et al. (1998) suggested that in some cases HEPA filters are unnecessary and it was better to have a high-efficiency filter with outside air to lower the pressure drop and increase system efficiency.[60] Another study by the same research group indicated that for removing 1 μm and larger common spores, 90% efficient filters were as effective as HEPA filters.[61] Miller-Leiden et al. (1996) showed that for removing aerosol with 0.7 and 1.3 μm non-HEPA units was as effective as HEPA units.[62]

Stephens and colleagues (2013) offered a method by modifying the Wells–Riley approach to calculate the risk related to airborne infectious diseases in a hypothetical hospital waiting room. The analysis showed that having HVAC filtration greatly reduced the risk of HAIs, with the lowest cost of operation depending on climate and building. MERV13–16 were more effective in reducing HAIs compared to MERV 11.[63] Another simulation of airflow and contaminant transport was used to analyze the decay of TB concentration over time. Between various filtration setups, no significant difference was found between removal rates of MERV 15 and HEPA filters.[20] The main disadvantages of HEPA filters are that they are costly in terms of maintenance, and also in terms of electricity requirements, as they have high pressure drop which increases the fan energy consumption.[60]

Filtration in Operating Room (OR)

High efficiency filters have been widely used in operating rooms to reduce the risk of surgical site infection.[64,65] The SARS-CoV-2 pandemic has brought to light an abundance of relevant publications pertaining to filtration in the OR. For example, during an HVAC system operations audit of 20 operating rooms in 10 different hospitals in the Hellenic Republic, the most common problems were poor maintenance of filters and nonexistent of air exhaust filters.[66] Another study compared freestanding HEPA filter units placed inside the OR to a novel portable anteroom system with HEPA unit (PAS-HEPA) that is placed outside the OR, near the main entrance.[67] It was found that with the PAS-HEPA system, experimental smoke plumes were observed moving downward and toward the main entry door. However, when freestanding HEPA units were placed on the patient bed inside the OR, smoke plumes were moving vertically upward toward the breathing zone of the OR personnel. These findings suggest that PAS-HEPA systems can be adopted when coping with surgical patients during an airborne pandemic. Al-waked (2010) investigated the steady state operation of a laminar flow system (LAF) using HEPA filters that was centrally located in the OR ceiling, using computational fluid dynamics (CFD), and results indicated their effectiveness in minimizing the spread of contaminants and providing thermal comfort.[68]

One major difference between the OR design and that for the AIIR is the use of laminar airflow systems. The laminar flow system provides a unidirectional flow,[36] and it must not be confused with cleaned air.[69] The HEPA filtration plays an outstanding role in the efficiency of laminar flow systems, as it assures the cleanliness of the supply air prior to pouring it on the OR bed.[70−72] The laminar air flow combined with HEPA filters can bring the contamination level down and is recommended for ORs and immunocompromised patient rooms.[37,73−78] It must be noted that a true laminar flow system is not achievable, as hospital personnel and equipment movements often disturb airflow patterns.[7]

In a study by Sheretz et al. (1987), the use of whole-wall HEPA filtration units with horizontal laminar flow (LAF) in patient rooms reduced the number of Aspergillus organisms in the air to 0.009 colony-forming units/m3, which was significantly lower than in all other areas of the hospital.[78] However, Sadrizadeh et al. (2016) showed that horizontal laminar flow systems, even when equipped with HEPA filters, lose efficiency due to the movement of medical staff and equipment.[79]

The Surgical Infection Society recently published guidance on delivering surgical care to SARS-CoV-2 patients. They recommended the use of a negatively pressurized OR, should one be available. Facilities that lack negative pressure ORs shall equip a room with a HEPA filter for surgical purposes. Also, the guidance requires a geographical separation (including the supply room) between the OR dedicated to COVID-19 patients and the rest of the surgical suite. No particular mention of air recirculation was found in the guidance.[80]

Filtration in AIIRs

Research on 2519 burn patients provided strong evidence that single-bed isolation rooms in combination with air filtration substantially reduced the incidence of infection with Gram-negative bacteria.[53] Another study measured the efficiency of isolation rooms with and without anteroom and with and without care provider traffic. Two micrometer aerosol particles that were dyed with fluorescent dye were released inside three isolation rooms that were equipped with HEPA filters. The negative pressure and HEPA filtration was effective in containing more than 99% of the aerosols.[81] The experiments were repeated for different provider traffic levels (i.e., no pass, one pass, and three pass) and results revealed over 99.7% containment of aerosols under all three conditions, providing a 5 min stay inside the AII unit. This is particularly important as it relates to providing care to the COVID-19 patients in AIIRs.

Marier and colleagues (1993) investigated a novel filtration–ventilation system to recirculate AIIR air through an Ultra-Low Penetration Air (99.9995% efficiency for 0.12 μm particles) filter and return the clean air into the adjacent anteroom. This system partially contributed to establishing the negative pressurization in the AIIR. The concentration of particles reduced to 30% of the baseline (fan off) when the system worked on the low mode, and up to 2% of the baseline for the high mode.[82]

Filtration in Protective Environment (PE)

Bone marrow transplant (BMT) units in the University of Minnesota hospital, built in 1986, had three different filters: 1, rough filter 40% (ASHRAE 52–76); 2, 95% bag filter (ASHRAE 52–76); and 3, HEPA final filters.[83] Other rooms in the hospital had the same filters, with the exception of HEPA filtration. By mapping particle concentrations to the level of filtration and occupancy level in the room, it was found that the number of small particles (0.3–0.5 μm) was a function of the filtration, and that the number of large particles (>5 μm) was a function of the occupancy level. Compared to the rooms without HEPA filters, the count of small particles (0.3–0.5 μm) was substantially lower for BMT units equipped with HEPA filter. Hahn et al. (2002) also showed that HEPA filters made a significant reduction in the number of HAIs in highly immunocompromised patients.[77] The incident of nosocomial Aspergillus infection was 10 times lower when the BMT patient was placed in a room with HEPA filtration and laminar airflow.[78] Passweg et al. (1998) also found that HEPA filters reduced contamination and increased the survival of BMT and leukemia patients.[54]

During a four-year period, in two hematology wards and a BMT unit, there were 64 cases of invasive nosocomial Aspergillus. The rooms were equipped with HEPA filters and/or laminar air flow (LAF) systems. Air and surface samples were collected to measure the level of fungal contamination bimonthly during this period. Results showed no fungi in air or surface in the rooms equipped with HEPA-LAF,[73] consistent with the findings of other investigators.[78,84,85]Table presents a summary of publications on air filtration with a brief discussion of the employed method.

Table 4

Summary of Select Publications on Filtration in Healthcare Settingsa

reference	quality level	setting	methodology	remarks
Shirani et al. (1986)[55]	2	burn units	Observational study of 318 patients	Significant improvement was observed in the cohort of patients admitted to a renovated unit. The renovation consisted of adding handwashing sinks, partitions to provide individual rooms, and HEPA filters on the air supply.
Sheretz et al. (1987)[78]	3	BMT	Observational study of 113 patients	Placing patients in a room with a whole-wall HEPA filtration unit reduced the risk of nosocomial aspergillus up to 10 times.
Barnes and Rogers (1989)[84]	2	BMT	Observational study of 19 children	The introduction of laminar airflow plus HEPA filtration terminated the outbreak of invasive pulmonary aspergillosis.
Marier et al. (1993)[82]	3	AIIR	Experiments in controlled environment	The combination of UV lights and Ultra Low Particulate Air (ULPA) filters efficiently removed particles from the air.
McManus et al. (1994)[53]	2	Burn units	Observational study of 2519 patients over ten years	Isolation of burn patients in separate rooms equipped with new filters reduced mortality ratio to one-third of predicted ratio. Authors attributed the improvements to the use of single-bed rooms, rather than the filtration system. However, these two effects were not decoupled.
Miller-Leiden et al. (1996)[62]	2	Test Chamber	Experiments in controlled environment	Ceiling mounted filters reduced the concentration of synthetic aerosols tracer particles by 90%. Non-HEPA filters were as effective as HEPA filters.
Passweg et al. (1998)[54]	2	PE	Observational study of 5065 patients	LAF+HEPA filtration significantly reduced the mortality rate in the first 100 days. The combination of LAF and HEPA filtration was effective. The influence of LAF from HEPA filtration was not decoupled.
Cornet et al. (1999)[85]	2	PE	1047 prospective air sampling during 2-year period	Efficiency of HEPA filtration and LAF+HEPA in preventing Aspergillus contamination was assessed. It was shown that HEPA filtration alone may not be sufficient under sever release due to construction/renovation activities.
Alberti et al. (2001)[73]	2	BMT	prospective study of 3100 air and 9800 surface samples	Fungal contamination was never found in air and on surfaces of rooms with HEPA+LAF. Separate effects of HEPA and LAF were not decoupled.
Hahn et al. (2002)[77]	2	PE	Retrospective cohort study of 90 patients	An outbreak of invasive aspergillosis was documented in a hematologic oncology unit with no HEPA filter. The contamination source was in determined in the non-BMT wing of the setting.
Olmsted et al. (2008)[67]	3	OR	Experiments in controlled environment	Using a freestanding HEPA unit inside the OR resulted in a surge of synthetic particles into the sterile zone. Using it outside of the room could effectively remove the particles.
Johnson et al. (2009)[81]	3	AIIRs	Experiments in controlled environment	The effect of HEPA filtered air recirculation with AIIR was assessed. In the presence of abundant particles, some might escape through the HEPA unit.
Stephens et al. (2013)[63]	4	Waiting Room	Numerical	The Well-Riley equation was modified to include the removal efficiency of filters. The influenza infection risk was not mitigated using a filter rating higher than MERV 13.
Emmerich et al. (2013)[20]	4	General Ward	Numerical	A well-mixed condition was assumed. Concentration of TB was reduced by 3 orders of magnitude when HEPA filters were used. The use of HEPA filters led to significant decrease in contaminant concentrations compared to MERV 15.

Data are presented in chronological order.

Portable HEPA Filters

HEPA filters can reduce the fungal spore and pathogen counts.[73,86,87] HEPA filters are usually placed in the HVAC system.[88] It is better to locate the HEPA filter at the point of air discharge to the room rather than at the discharge of the fan.[89−91] This is because the ductwork system from the fan discharge to the room outlet could be contaminated due to maintenance or construction.[92] Portable HEPA units can be used if this practice is not feasible. Portable HEPA filters are capable of increasing the effective ACH, removing particles, and reducing the risk of transmission in general wards when there are not enough isolation wards as is the case with the COVID-19 pandemic[8,39,89,93−95] (Table ).

Table 5

Summary of Publications on the Use of Portable HEPA Filter Units in Healthcare Facilitiesa

reference	quality level	setting	methodology	remarks
Rutala et al. (1995)[89]	3	AIIR	Experiments in controlled environment	Portable HEPA units were effective in removing aerosols from the room both as a standalone mechanism and as a supplement to existing ventilation system. Location of the unit was not found important.
Miller-Leiden et al. (1996)[62]	3	Test chamber	Experiments in controlled environment	Efficiency of portable units was lower than the ceiling mounted filters. The portable units performed much better at high ventilation rates when short-circuiting was avoided.
Rebmann (2005)[39]	4	AIIR	Algorithm development	The use of portable filter units was recommended in the event of emergency patient isolation.
Boswell (2006)[92]	2	Patient room	Agar plate data collection	CFUs of MSRA significantly reduced when portable HEPA units were used. Ventilation rate of the unit did not seem to have a considerable effect.
Qian et al. (2007)[93]	3	Test chamber	Experiments + simulation	A minimum of <5.6 ACH from the portable unit was required to clean air with no other means of room ventilation. With the room HVAC system on, this number was reduced to 2 ACH.
Bergeron et al. (2007)[107]	3	OR + pediatric hematology room	Experiments in simulated surgical conditions	Particle decay time reduced from 12 min to <2 min when the portable unit was added. Concentrations of airborne mesophilic were halved.
Abdul Salam et al. (2010)[88]	2	acute tertiary-care	retrospective study of 134 cases	The incidence rate of invasive Aspergillus was dropped by a factor of 2 upon addition of a portable HEPA unit.
Rao et al. (2020)[98]	2	Pediatric Hospital	Non-randomized study of 562 patients	The use of air purifier significantly reduced hospitalization time and the rate of using noninvasive ventilation techniques.

Data are presented in chronological order.

Miller et al. (1996) investigated the effectiveness of portable air filters (PAFs) and ceiling mounted air filters (CMAFs) for controlling TB exposure. CMAFs were shown to be more advantageous over PAFs due to their elevated inlet and the buoyant nature of the droplet nuclei.[62] Qian and colleagues tested the performance of portable HEPA filters, such as their effect on the airflow pattern, using both smoke visualization and computer simulation. They found that having portable HEPA filters with strong air supply could bring a global airflow mixing by interacting with the airflow pattern.[93] One study found, on average, a 75% reduction in MRSA surface contamination when portable HEPA filter was placed inside three different patient rooms.[96] While the portable unit did make a significant change in overall airflow patterns, the final outcome (i.e., surface decontamination) did not change by changing the location of the HEPA unit. Casagrande and Piller (2020) demonstrated that the change in the global patterns of air movement due to a portable HEPA unit can be positively harnessed. By placing the portable HEPA unit next to the instrumentation table in the OR, an invisible air curtain was created around the instrument table that reduced the risk of instrument contamination up to 75%.[97] Another investigation assessed the impact of 48 portable HEPA filter units on the invasive Aspergillosis (IA) incidence. Portable HEPA filters nearly halved the IA incidence ratio. The costs of HEPA filtration were negligible compared to the significant decrease in the rates of IA infection.[88] Nevertheless, portable HEPA filters can be noisy, and bring physical obstruction for airflow in the room.[39] A more recent study assessed the performance of a portable air purifier (filter efficiency not specified) in a pediatric hospital setting. A total of 562 patients diagnosed with infectious or noninfectious respiratory distress were included in this study in pre- and post-cohorts. Results showed that the rate of using invasive patient ventilation (i.e., intubation) was reduced 30%, and the average hospitalization time declined from 0.7 days to 0.4 days upon placing the portable air purifier in the rooms,[98] though these differences did not achieve statistical significance. Further, clinical details were not provided to allow a better understanding of potential differences in specific clinical outcomes such as healthcare associated pneumonias, so it is difficult to understand these changes.

Other Air Treatment Methods

While filtration is critical in maintaining the quality of hospital air, filters themselves can shelter viable organisms and hence sometimes support their growth.[99−101] Therefore, the device itself can become the source of contamination.[102−107] In an investigation of three sealants with different active antimicrobial compounds, only two of them reduced the fungal regrowth.[108] Some of the air microorganisms are compatible with the antimicrobial agents applied on them.[109] On the other hand, Foarde et al. (2001) found that filters did not become the source of microbial growth as long as polyacrylate copolymer antimicrobial agents containing zinc oxide and borates were used under normal conditions.[110] It was also reported that antimicrobial agents were not effective in curbing fungal reproduction on dust-loaded filters, while chemically coated filters had slightly higher than 60% efficiency.[111] This area of research has remained fairly controversial where there are not enough studies to support (or impugn) the use of antimicrobial agents on air filters.[20,112] Specific to the coronavirus outbreak, rigorous research must be conducted to characterize the viability and reproduction of the virus on the filters. These data are particularly critical when deciding on protocols for replacing, storing, and disposing of used filters. Bergeron and colleagues (2007) designed a novel mobile air treatment unit to both remove airborne particles and destruct microorganism by use of electrostatic capture and nonthermal plasma reactors.[107] Nonthermal plasma reactors ensure that no microorganisms were stored in the system and eventually released into the air. Their efficiency for removing biological airborne contamination at the speed of 1.5 m/s was reported greater than 99% for a wide range of microorganisms.[113]

The combination of filtration and disinfection has been recommended by CDC, NIOSH, and ASHRAE as a complementary air cleaning approach to the room air conditioning system.[12,15,114] For example, the combination of ultraviolet germicidal irradiation (UVGI) lamps and filtration was most effective in removing and killing pathogens like Aspergillus spp. and Staphylococci.[91,115] Ryan et al. found that installation of UVGI in the HVAC systems equipped with 95% filter efficiency helped reduce ventilator-associated pneumonia and tracheal colonization.[116] Another more recent work studied the effectiveness of UVGI when used in ventilation ducts. Results showed the UVGI inactivated nearly all the microorganisms in the ductwork when the air velocity was 3.0 m/s. However, an increase in air velocity (6.0 m/s) led to insufficient exposure of bacteria to irradiation, and thus only 80% inactivation was found for S. enterica and S. epidermidis.[117] Interested readers may find further useful discussions in a review article published by Memarzadeh and colleagues on the application of the UVGI technology in healthcare settings.[118]

Recirculation in Healthcare Facilities

Waiting for a contaminant to decay naturally in the building is a slow process. Hence, engineering alternatives like mixing indoor air with the outside air and recirculating it back to the room and pressurization control are utilized. There are two types of general air systems: (1) single pass and (2) recirculating.[119] The single pass air system supplies air from the outdoors, and after passing through the room air it is then exhausted 100% back to the outdoor environment. This system does not, however, take advantage of the embodied energy required for heating/cooling of discharged air. In the recirculating system, however, a limited portion of the air is exhausted to the outdoors and replaced by fresh air.[120] Recirculating systems are typically designed for significantly larger flow rates compared to single pass systems as they utilize the power of contaminant dilution. Recirculation could be used as long as air is filtered through high efficiency filters or is cleaned by disinfection techniques.[121] Recirculation of air can be problematic under two conditions: (1) when it leads to the delivery of insufficient fresh air, and consequently excessive CO2 exposure, or (2) when the return air is not properly filtered.[122] For most patient-care areas, there are two filters in central HVAC systems. The outside air is filtered through the first bank filter to remove larger particles (1–5 μm). The incoming air then mixes with the recirculated air and is reconditioned for temperature and humidity and eventually passes through the second filter in the air handling unit.[35,123]

Ulrich (1974) studied two neurosurgical operating rooms at St. Mary’s Hospital in Rochester, Minnesota, during surgical procedures. Air and surface samples were taken to determine bacterial counts. Results for both rooms demonstrated that recirculation of air along with filtration reduced the number of bacteria inside the OR when compared to 100% fresh air.[124] Woods et al. (1986) studied two operating rooms in Mary Medical Center in Ames, Iowa, and data were recorded from two existing HVAC settings. In the first OR, the system performed 12 air changes per hour with 100% outdoor air and only one filter. In the second operating room, the system performed 17 air changes per hour with 20% from the outdoor (i.e., 3.5 ACH) and two filters. Although the percentage of outdoor air in the second operating room was lower than the first room, computer simulations and field measurements showed a reduction in particulate matter concentrations, indicating the usefulness of air recirculation and filtration.[125]

Research has shown that air recirculation should only be used in the hospital setting in conjunction with adequate air filtration. The distribution of particles were studied in a patient room of a Brazilian hospital that was ventilated with a split system that recirculated air inside the room with outdoor air, but without filtration. Results showed that small particles followed the path of the air and remained suspended in the room due to air recirculation inside the room. Hence, pure recirculation can increase the risk of contamination due to random particle movements, and it may also result in significant particle deposition onto surfaces.[126] The placement of inlet and outlet filters is also important, as the HVAC system can cause different recirculation patterns in the room.[127,128]

Practical Implications

There is substantial evidence (Level 2 publications) that contaminated air can result in disease spread, and that the combination of air filtration and recirculation can reduce this risk. Observational and animal studies suggest that air recirculation alone may result in the airborne transmission of pathogens.[129−131] The experimental setup is mostly designed for extreme cases to prove the airborne route, which is far from realistic in the healthcare setting. There are a few outstanding findings from the literature that can be used to minimize the adverse impact of the SARS-CoV-2 virus on medical personnel who spend many hours of their time inside the hospital as well as reduce risk of nosocomial infections.

The combination of HEPA filtration and air recirculation has been shown to be extremely effective in many space functions such as ORs, AIIRs, and PEs. These systems clean air by simultaneously removing (i.e., filtering) and diluting (through recirculation) contaminants from the space. They have shown significant reduction in the number of bacterial colonies, and surrogate particulate matters previous studies, and they are expected to show a similar efficiency for the COVID-19 pandemic. The role of air distribution systems in buildings during an outbreak has been the subject of previous research.[132,133]A finding that is worthy of reiteration is from Streifel and colleagues (1995) that showed HEPA filters are remarkably efficient in capturing submicron particles.[83] The SARS-CoV-2 virus is shown to have a diameter less than 1 μm. This further suggests that HEPA filters will be effective tools to mitigate hospital spread, as noted in a recent review on the potential airborne route of the SARS-CoV-2 transmission in hospitals.[134]

The coronavirus outbreak has severely hit the healthcare settings. Notably, in hospitals in the epicenter of the pandemic during the spring of 2020 (New York, Connecticut, and Boston), the surge capacity exceeded standard room configurations and SARS-CoV-2 positive patients were placed in general patient wards and converted ICUs without negative pressure capacity. These spaces likely frequently lack sufficient filtration requirements, ventilation rates, and negative pressurization capacity. Therefore, one short-term solution seems to be substituting the existing filter in those rooms with higher efficiency filters (i.e., HEPA filters). Though this seems like a logical decision at the first glance, it may result in unfavorable outcomes such as increasing the pressure drop when replacing a lower rate filter with HEPA filters.[29,135] When the pressure drop increases and the fan speed remains constant, the amount of air supplied to the space will decrease. This can result in an imbalanced system that can increase the distribution to spread disease.

Portable HEPA filter systems have been shown to be another effective tool in reducing viral load. Another innovative design is to place a portable HEPA filter that discharges air into a portable plastic anteroom right at the entrance.[82] Such a system, also recommended by the ASHRAE epidemic task force, can clean the air inside the patient room, while creating negative pressure with respect to the plastic anteroom. Therefore, both virus removal and containment may be achieved.