eAppendix 1. Literature Search Strategies
eFigure. PRISMA Flow Chart of Article Selection
eAppendix 2. Ultraviolet Germicidal Irradiation (UVGI)
eAppendix 3. Vaporized Hydrogen Peroxide (VHP)
eAppendix 4. Ethylene Oxide (EtO)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Schumm MA, Hadaya JE, Mody N, Myers BA, Maggard-Gibbons M. Filtering Facepiece Respirator (N95 Respirator) Reprocessing: A Systematic Review. JAMA. 2021;325(13):1296–1317. doi:10.1001/jama.2021.2531
What methods of filtering facepiece respirator decontamination are effective and feasible?
Five decontamination processes and 42 studies were reviewed. Ultraviolet germicidal irradiation, moist heat, and microwave-generated steam processing were effective for pathogen removal, preserved respirator filtration, and had short treatment times and readily available equipment. Vaporized hydrogen peroxide is a suitable alternative with longer decontamination durations and is more expensive. Ethylene oxide may leave toxic residues and is less easily implemented.
Ultraviolet germicidal irradiation, moist heat, and microwave-generated steam processing of filtering facepiece respirators are effective means for decontamination and are simple to implement.
The COVID-19 pandemic has resulted in a persistent shortage of personal protective equipment; therefore, a need exists for hospitals to reprocess filtering facepiece respirators (FFRs), such as N95 respirators.
To perform a systematic review to evaluate the evidence on effectiveness and feasibility of different processes used for decontaminating N95 respirators.
A search of PubMed and EMBASE (through January 31, 2021) was completed for 5 types of respirator-decontaminating processes including UV irradiation, vaporized hydrogen peroxide, moist-heat incubation, microwave-generated steam, and ethylene oxide. Data were abstracted on process method, pathogen removal, mask filtration efficiency, facial fit, user safety, and processing capability.
Forty-two studies were included that examined 65 total types of masks. All were laboratory studies (no clinical trials), and 2 evaluated respirator performance and fit with actual clinical use of N95 respirators. Twenty-seven evaluated UV germicidal irradiation, 19 vaporized hydrogen peroxide, 9 moist-heat incubation, 10 microwave-generated steam, and 7 ethylene oxide. Forty-three types of N95 respirators were treated with UV irradiation. Doses of 1 to 2 J/cm2 effectively sterilized most pathogens on N95 respirators (>103 reduction in influenza virus [4 studies], MS2 bacteriophage [3 studies], Bacillus spores [2 studies], Escherichia virus MS2 [1 study], vesicular stomatitis virus [1 study], and Middle East respiratory syndrome virus/SARS-CoV-1 [1 study]) without degrading respirator components. Doses higher than 1.5 to 2 J/cm2 may be needed based on 2 studies demonstrating greater than 103 reduction in SARS-CoV-2. Vaporized hydrogen peroxide eradicated the pathogen in all 7 efficacy studies (>104 reduction in SARS-CoV-2 [3 studies] and >106 reduction of Bacillus and Geobacillus stearothermophilus spores [4 studies]). Pressurized chamber systems with higher concentrations of hydrogen peroxide caused FFR damage (6 studies), while open-room systems did not degrade respirator components. Moist heat effectively reduced SARS-CoV-2 (2 studies), influenza virus by greater than 104 (2 studies), vesicular stomatitis virus (1 study), and Escherichia coli (1 study) and preserved filtration efficiency and facial fit for 11 N95 respirators using preheated containers/chambers at 60 °C to 85 °C (5 studies); however, diminished filtration performance was seen for the Caron incubator. Microwave-generated steam (1100-W to 1800-W devices; 40 seconds to 3 minutes) effectively reduced pathogens by greater than 103 (influenza virus [2 studies], MS2 bacteriophage [3 studies], and Staphylococcus aureus [1 study]) and maintained filtration performance in 10 N95 respirators; however, damage was noted in least 1 respirator type in 4 studies. In 6 studies, ethylene oxide preserved respirator components in 16 N95 respirator types but left residual carcinogenic by-product (1 study).
Conclusions and Relevance
Ultraviolet germicidal irradiation, vaporized hydrogen peroxide, moist heat, and microwave-generated steam processing effectively sterilized N95 respirators and retained filtration performance. Ultraviolet irradiation and vaporized hydrogen peroxide damaged respirators the least. More research is needed on decontamination effectiveness for SARS-CoV-2 because few studies specifically examined this pathogen.
SARS-CoV-2 is an RNA virus that has a diameter of approximately 0.1 μm. It is transmitted by respiratory droplets and aerosols such that one of the most effective ways to prevent infection is to wear filtering facepiece respirators (FFRs). The commonly used FFRs are N95 respirators, devices that by design filter 95% or more of particles larger than 0.3 μm in diameter. In practice, these devices prevent the passage of 99.8% of particles larger than 0.1 μm.1 RNA viruses are inherently unstable, and transmission of viable SARS-CoV-2 can occur only if the virus is protected from the atmosphere by being contained within moist respiratory droplets or aerosols. The droplets are of various sizes, with large droplets being on the order of 5 μm or larger.2 Respiratory aerosols can transmit the virus great distances and have particle sizes of about 1 μm. All of these particle sizes are effectively filtered by N95 respirators, as long as they closely fit the contours of the wearer’s face, avoiding leaks around the filter material.
The COVID-19 pandemic initially resulted in a critical shortage of all forms of personal protective equipment, especially N95 respirators.3,4 Because of limitations of the FFR supply chain, N95 respirators remain scarce, with little hope of having an adequate supply while the pandemic lasts. Consequently, most health care institutions have resorted to reprocessing these devices that were intended to be used one time only. Centers for Disease Control and Prevention (CDC) guidance exists for FFR reprocessing in emergency conditions.5 However, that guidance is associated with few recommendations for how to reprocess these devices. Filtering facepiece respirators can fail if the reprocessing system cannot kill all the pathogens that accumulate within the mask material itself, if the filters are compromised and lose filtering efficiency, or if the mask elasticity is altered such that it no longer provides a tight fit and leaks air around the mask.
To date, several approaches have been tested for the reprocessing of N95 respirators and how they affect filtration efficiency and facial fit characteristics.6-8 They include UV germicidal irradiation (UVGI), vaporized hydrogen peroxide (VHP), moist-heat incubation (MHI), microwave-generated steam (MGS), and ethylene oxide. How to select the optimal method is uncertain. To determine which method is most clinically useful, we performed a systematic review to assess efficacy and feasibility of each reprocessing strategy.
This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.9 Two librarians developed a search strategy for FFRs (including N95s and respirators), sterilization, UVGI, VHP, MHI, MGS, and ethylene oxide to identify PubMed and EMBASE articles. Research reports were assessed between database inception and January 31, 2021.
Titles were reviewed independently by 2 individuals to determine if the study assessed FFR decontamination results using UVGI, VHP, MHI, MGS, or ethylene oxide. Studies evaluating decontamination of impregnated pathogens (including bacteria, virus, fungus, and spores) from FFRs and/or how this affects the mask’s filtration efficiency and facial fit were included. Efficacy of pathogen decontamination was determined by the reduction in pathogen load, with greater than 103 reduction after reprocessing considered efficacious, corresponding to greater than 99.9% inactivation of pathogens per US Food and Drug Administration guidelines.10 Respirator filtration performance was assessed by filtration efficiency (particle penetration filtered by the material and airflow resistance across the filter material). Facial fit was assessed with quantitative and/or qualitative fit test methods. (The fit factor obtained with a quantitative fit test calculates a ratio of the test agent concentration outside the respirator to the test agent concentration inside the respirator, with a range of 1-200 for N95 respirators, and >100 needed to pass). A comparator group was not required for inclusion.
Studies evaluating decontamination of pathogens on non-FFR surfaces, review articles, and editorials were excluded. Reference mining was performed. The gray literature and manufacturer websites were reviewed for cost data of equipment required for each decontamination method. Direct communication was made with manufacturers when data were not available online. The CDC website was assessed, which identified additional non–peer-reviewed (ie, gray) literature. The full search strategy is provided in eAppendix 1 in the Supplement.
The N95 respirator is one of the most common types of respirators used by health care practitioners. It is 1 of 9 types of disposable particulate FFRs and protects by filtering particles out of the air that is breathed.11 N95 respirators filter out at least 95% of airborne particles, such as bacteria or virus particles, while those filtering out at least 99% receive a 99 rating and those that filter at least 99.97% receive a 100 rating.12 Respirators are rated N (not resistant to oil), R (somewhat resistant to oil), and P (strongly resistant [oil proof]). When approved by the National Institute for Occupational Safety and Health (NIOSH), respirators are marked with the manufacturer’s name, part number, protection provided by the filter, and the acronym “NIOSH.”
Because of heterogeneity in study designs and within-study variations of the processes tested and outcomes measures, data pooling was not possible; hence, the synthesis is narrative. Data were dual abstracted for decontamination process, treatment system, and 3 indicators of quality of processing, including effectiveness of removal of pathogen, physical integrity/performance of the mask (filtration efficiency and/or facial fit), and safety for users. Throughput (number of FFRs processed per cycle) and processing capability were reported. If only UV intensity and time of exposure were specified, UV dose was calculated using the equation UV dose (J/cm2) = intensity (W/cm2) × time (s).
The literature search identified 238 articles, with an additional 8 identified through reference mining. After screening, 42 studies were included (eFigure in the Supplement); all were laboratory studies (no clinical trials), and 2 evaluated respirator performance and fit with actual clinical use of FFRs. Of the 42 studies, 26 included evaluations of the following pathogens/organisms: SARS-CoV-2 (5 studies), influenza virus (4 studies), MS2 bacteriophage (8 studies), Escherichia coli (2 studies), Staphylococcus aureus (1 study), Mycobacterium smegmatis (1 study), Escherichia virus MS2 (1 study), vesicular stomatitis virus (1 study), and Bacillus and Geobacillus stearothermophilus spores (6 studies). Overall, 27 studies evaluated UVGI decontamination, 19 evaluated VHP, 9 evaluated MHI, 10 evaluated MGS, and 7 evaluated ethylene oxide.
Sterilization using UVGI depends on total power of UV energy delivered from specialized bulbs or lamps that illuminate a confined space such a box, biosafety cabinet, or entire room. N95 respirators have several layers of material that the UV energy must penetrate. Ultraviolet germicidal irradiation causes radiolytic stress that disrupts a microorganism’s DNA but may degrade polymers commonly used in disposable N95 respirators.
The effectiveness of UVGI to sterilize FFRs impregnated with infectious pathogens was examined in 17 of the 27 studies evaluating 31 different commercially available models of N95 respirators. Substantial heterogeneity existed among studies in systems used to deliver UVGI, dose of UV light applied, and FFR models and pathogens studied. Ultraviolet germicidal irradiation energy of 0.03 to 7.2 J/cm2 was applied with exposure times ranging from 1 to 300 minutes. Overall, doses exceeding 1 J/cm2 effectively decontaminated FFRs as measured by reduction in influenza virus (4 studies), MS2 bacteriophage (3 studies), Bacillus spores (2 studies), Escherichia virus MS2 (1 study), vesicular stomatitis virus (1 study), and Middle East respiratory syndrome virus and SARS-CoV-1 (1 study) (Table 1).
Higher doses of UVGI energy were needed to decontaminate FFRs for SARS-CoV-2 compared with other viruses such as influenza virus and MS2 bacteriophage. Two studies showed that 1.50 to 1.98 J/cm2 were required for SARS-CoV-2 decontamination of FFRs,13,14 whereas 3 other studies did not find that UVGI effectively decontaminated FFRs.15-17 In a study requiring 1.98 J/cm2, the power of the UV lamp was lower than in most other studies assessing pathogen decontamination (median, 2.1 [interquartile range, 0.66-13.8] mW/cm2) and may have led to decreased penetration of the UV rays through the N95 respirator’s material.13 No other studies evaluated the specific FFR model that was examined in that study (the AOSafety N9504C). In another study, UV box and whole-room decontamination treatment with an unspecified level of UV power or dose did not effectively decontaminate 3 different N95 respirators types when Pseudomonas phage φ6 and MS2 bacteriophage were suspended on the outer top, outer edge, and inner surface of the respirator facepiece.15 Although UV dose was not specified, UV-C colorimetric indicators demonstrated delivery of a dose sufficient to reduce Clostridioides difficile spores. Virus was suspended in liquid and spread over a small surface area, which may have led to differences in efficacy compared with droplet and aerosol deposition. The efficacy of UV light increases when an inoculum is spread over a larger surface area.18 In another study assessing SARS-CoV-2 decontamination of 3 N95 respirator models,16 0.63 J/cm2 led to effective virus reduction in only 1 of the models (the 3M 1870+). However, high titers of SARS-CoV-2 were directly infiltrated into the N95 respirators in this study beyond what may occur in a more realistic exposure scenario, such as from a single cough,19 and only a single decontamination experiment was performed. Certain models of N95 respirators in which UVGI was found to be less effective include the Gerson 1730, the Sperian HC-NB095 (now manufactured by Honeywell), and the Dentec Safety AD2N95A20,21; the 3M VFlex 1805 and the Precept 65-339521; and 3M 1860S, the Moldex-Metric 1517, and the Kimberly-Clark 46727.15
The potential for UVGI to damage FFRs and diminish their filter performance and facial fit was examined in 23 studies and 30 different commercially available models of N95 respirators. Of these, 1 study demonstrated diminished strength of the respirator materials from UVGI (Table 1). However, this was after very high doses (>120 J/cm2) beyond what is used in clinical practice.22 Overall, N95 respirator fit characteristics and filtration performance were not affected in up to 20 decontamination cycles with high cumulative doses,6,16,17,21-27 with the exception of 3 recent studies in which filter efficiency decayed to less than 95% after 1 cycle and 20 cycles with an unmeasured UV dose,28,29 and the 3M 1860 and 3M 1860S filter performance degraded to 70% to 80% after 9 repeated cycles using whole-room decontamination.30 In the study applying 20 cycles of UV light,28 UVGI was applied to melt-blown fabric samples (the most critical filtering layer within the N95 respirator that allows air to pass through while filtering 95% of microbes) and was thus unprotected by other respirator layers.
Commercially available UVGI lamps or bulbs can be placed in chambers or cabinets to construct decontamination devices used to reprocess 4 to 40 masks per 5- to 10-minute cycle, corresponding to 1500 to 5000 masks per day, at a cost of $750 to $1500.31,32 Because UVGI devices and laboratory hoods are ubiquitous in many hospitals, this approach is feasible and potentially inexpensive.32,33 While commercial UV-C bulbs are relatively inexpensive ($40-$285),34 laboratory hoods and commercial UVGI whole-room decontamination systems are more costly ($7000-$21 00035,36 and $25 000-$40 000,37 respectively). Recommendations for how to select N95 respirators amenable to UV treatment, the UV sterilization device, and toxicity information are available in eAppendix 2 in the Supplement.
Vaporized hydrogen peroxide is a sterilant for heat-sensitive medical devices and equipment.38 Its use as a gaseous disinfectant circulated in an enclosed space or open room is safe given its low toxicity and residual gas vapor decomposition into water vapor and oxygen.
Nineteen studies assessed outcomes on 30 N95 respirators using VHP (Table 2). There was heterogeneity among studies in the reprocessing systems used (open-room vs closed chambers), in VHP concentration, in FFR models, and in pathogens analyzed. These studies examined commercial VHP generators that converted liquid hydrogen peroxide into vapor and, following hydrogen peroxide vapor treatment, required an aeration phase to safely eliminate toxic residual chemicals. STERRAD H2O2 gas plasma sterilizer systems (Advanced Sterilization Products) were used in 6 studies, while 10 studies used open-room decontamination systems (Bioquell Room Bio-Decontamination Service Clarus R hydrogen peroxide vapor generator, Bioquell Z system, Steris Life Sciences VHP Victory system, or Radiant Innovations Satej Plus VHP generator). The STERRAD system uses a 100-L (0.1-m3) chamber for decontamination, while the open-room generators that are designed for large-scale decontaminations were placed in 33-m3 to 111-m3 rooms with N95 respirators either hung across on a string or suspended on racks. Cycle time varied from 28 minutes to 300 minutes, with variable aeration times up to 18 hours. Total treatment times ranged from 1 hour to 19 hours. The STERRAD systems were run at 45 °C to 55 °C for 55 minutes. In 3 recent studies, VHP chamber decontamination was evaluated using the Steris V-PRO maX sterilization system39,40 and a Panasonic incubator with VHP generation capabilities.13 Cycle times for these 2 systems varied from 28 minutes to 290 minutes, respectively. The concentration of VHP ranged from 7% to 59% based on the device (7%-8% for the Radiant Satej Plus; 30%-35% for the Bioquell Clarus R and Clarus C VHP generator and Z vaporizer chamber and the Steris ARD and VHP Victory; and 58%-59% for the STERRAD NX and 100S and the Steris V-PRO maX).
Overall, VHP effectively decontaminated FFRs as measured by greater than 106 reduction of Bacillus and G stearothermophilus spores (indicating complete inactivation) (4 studies), inactivation of vesicular stomatitis virus (1 study), and inactivation of SARS-CoV-2 (3 studies) (Table 2). In a 2016 report of the 3M 1860 with a Bioquell Clarus C VHP chamber generator, an optimal 480-minute VHP decontamination process was developed.41 Total exposure time was 8 hours, and the system could treat more than 50 N95 respirators per cycle. Fifty cycles of VHP treatment demonstrated complete inactivation of G stearothermophilus spores after each cycle while preserving filter performance. Additionally, fit was unaffected for up to 20 VHP cycles when tested on manikins. However, more than 30 cycles resulted in elastic strap degradation when stretched. This finding contrasted with a recent study using the Battelle Critical Care Decontamination System in which there was no effect on strap integrity after 20 cycles.42 The viricidal efficiency of VHP using a 130-minute decontamination process was confirmed by demonstrating complete eradication of E coli, M smegmatis, and G stearothermophilus spores without signs of physical degradation after 15 cycles.43 Three recent studies assessed SARS-CoV-2 decontamination of N95 respirators. In 2 studies, rapid virucidal reduction in virus was demonstrated while preserving filtration performance and fit after 2 cycles of treatment.13,16 In the other study, VHP led to complete inactivation of SARS-CoV-2, although an extended 5-hour cycle time was required, compared with 1 hour for vesicular stomatitis virus.27 Results regarding toxicity are available in eAppendix 3 in the Supplement.
Open-room systems, the Panasonic incubator, and the Steris V-PRO maX system resulted in better FFR filter efficiency and facial fit than did the STERRAD chamber system. In the 6 studies investigating physical appearance and filter performance after STERRAD system treatments, changes to the nosebands or significant reduction in filtration performance were observed (Table 2). When this system was examined with a 3-cycle treatment of 6 N95 respirator models, mean penetration levels increased above the NIOSH certification limit of 5% in 4 of 6 mask models. High exposure from the H2O2 gas plasma processing conditions may have resulted in reduced filter performance.8,27,30 There were no physical changes or reduction in filtration ability when open-room VHP systems were used8,26,27,43 or when FFRs were suspended or placed in exposure chambers.16,29,30,41 Filtration performance and fit was not reduced after 2 rounds of decontamination and 2 hours of wear using the Panasonic incubator13 or after 10 cycles with the Steris V-PRO maX sterilization system.30,39 However, one recent study demonstrated the combination of prolonged clinical FFR use and chamber VHP decontamination to be associated with early mechanical and fit test failure,40 while another study demonstrated fit test failure with several FFR models after repeated decontamination cycles (Halyard FLUIDSHIELD 46727) and when the mask was worn prior to decontamination (3M 1870).44
High-capacity throughput is currently performed using dedicated VHP systems for whole-room decontamination (>1000 N95 respirators per cycle43,45,46), although prolonged treatment and aeration times are required. The processing capability of VHP chamber systems may be limited by the volume of the chamber and the specific types of materials that FFRs are made of. Cellulose materials, such as cotton, absorb hydrogen peroxide and can cause the STERRAD device to stop because of low VHP concentrations within the chamber. This was observed in one study in which the sterilization cycle aborted when more than 6 N95 respirators were loaded into the device.47 However, the respirators in that study were made primarily of polyesters and polyolefins, so the reason the device aborted was not clear. These devices are expensive for both chamber and open-room systems ($40 000-$130 000).48,49
Heat inactivates viruses by denaturing the proteins involved in attachment and replication within a host cell. However, concerns exist regarding the deleterious effects dry heat may have on certain materials, including those that comprise N95 respirators. Moist heat, on the other hand, is more effective at killing microorganisms, it distributes homogeneously across the surface being sterilized, and the lower temperature is less likely to degrade materials.
Nine studies evaluated infectious and mask integrity outcomes after MHI on 15 contaminated N95 respirators (Table 3). Three studies used a Caron model 6010 laboratory incubator, 2 used preheated 6-L sealable containers filled with 1 L of water, 2 used a heated cabinet, and 2 used a preheated chamber. In 5 of the 7 studies, moist heat was generated at 60 °C to 70 °C, relative humidity of 50% to 90%, and incubation times lasting 15 minutes to 30 minutes, with outcomes evaluated after 1 to 3 treatment cycles. Moist-heat decontamination was also evaluated using heated cabinets at 70 °C to 90 °C, with 0% to 70% relative humidity, and a 60-minute incubation cycle for up to 15 cycles,27,50 and in another study, a preheated chamber at 85 °C, with 30% and 100% relative humidity, was evaluated using a 20-minute incubation cycle for up to 20 cycles.28
Of these studies, 1 reported infectious outcomes for N95 respirators contaminated with SARS-CoV-2 and E coli, 1 reported infectious outcomes for N95 respirators contaminated with SARS-CoV-2 and vesicular stomatitis virus, and 2 reported infectious outcomes for FFRs contaminated with influenza virus using a preheated 6-L water filled container. Moist-heat incubation effectively reduced SARS-CoV-2, E coli, and vesicular stomatitis virus after 60-minute incubation periods27,50 and effectively reduced influenza A(H5N1) and influenza A(H1N1) viruses to undetectable limits after 20-minute6 and 30-minute7 incubation periods, respectively. The infectious outcomes for contaminated straps were not reported.
No signs of damage were found when FFR filter efficiency and facial fit were tested after 1 cycle or 50 cycles of container/chamber incubation.6,7,28 Filter performance was not reduced following 10 or 15 cycles of treatment in a heated cabinet.26,50 Similarly, there was no reduction in facial fit and comfort after 10 cycles.26,50 In contrast, when FFRs were decontaminated in Caron incubators, 1 to 3 cycles of 15- to 30-minute treatment did not alter filter performance or fit testing, but significant physical degradation was observed with 3M 1870 N95 respirators. The 3M devices had partial separation of the inner foam nose cushion from the FFR body and melting of head straps.8,23,51 The 3M devices did not degrade when treated by preheated container incubatation.6 In the same study, one 30-minute incubation cycle resulted in significantly impaired fit in 2 of 6 N95 respirators (3M 8210 and Moldex-Metric 2200) tested; however, the quantitative fit factors (measuring particle leakage around the face seal) for these masks were still greater than 100, suggesting that the masks provided adequate protection against infection.23
Treatment with MHI uses readily available equipment, requires short exposure times, and can be easily implemented in most health care settings. The commercially available Caron incubator device for MHI treatment is expensive (approximately $9000-$23 000),52,53 the processing volume of its incubation chamber is small, long preheating times are required, and physical damage to the FFRs can occur.8,23
Steam treatment is a known method for inactivating viruses on surfaces. Microwave-generated steam is an alternative to standard steam treatment, as much of the energy formed from microwave radiofrequency can be absorbed by water, reducing the potential of damaging N95 respirator materials.54 A single N95 respirator is placed outer side down on top of a reservoir filled with 50 mL to 100 mL of room-temperature tap water and then placed on a rotating glass plate in the microwave to allow generated steam to distribute over the respirator surface. The operator uses a respirator holder placed above the water source to generate the steam needed to reduce the risk of degradation when using dry heat.54,55
Ten studies evaluated decontamination and filter performance for various pathogens following MGS decontamination of at least 10 FFR models (Table 4), but none studied SARS-CoV-2. A commercially available microwave was used in 9 of 10 studies. Devices ranged from 1100 W to 1800 W and were run at high power for various time ranging from 15 seconds to 3 minutes. In these studies, MGS effectively decontaminated FFRs for influenza virus (2 studies), MS2 bacteriophage (3 studies), and S aureus (1 study). Steam generated in 1100-W to 1800-W devices for 40 seconds to 3 minutes reduced MS2 bacteriophage, influenza A(H5N1), and influenza A (H1N1) virus counts by a factor of 103 to 104 and S aureus by 106.6,7,56-58 Microwave-generated steam was effective even when proteins such as those found in saliva or skin oil were present that tend to protect pathogens from decontamination.56,59 Contaminated FFR straps had a 105 reduction in viable MS2 bacteriophage following exposure to 2 minutes of MGS.57
Filter performance, respirator fit, and comfort were not affected by 1 to 3 cycles of MGS treatment in 5 studies6,8,23,51,58 and for up to 20 cycles in 1 study.57 However, FFR damage was observed in 4 studies. In 3 of these, 3M 1870 respirators had partial separation of the inner foam nose cushion from the FFR,8,23,51 and in 2 studies, the 3M 1870 and Kimberly-Clark PFR95-270 respirator head straps melted.8,51 Deformation of the FFR foam nose cushion was observed in a study evaluating an N95 respirator after a 2-minute MGS cycle using a 1250-W microwave oven, yet the elastic straps were unaffected.7 In a study using a commercially available 1250-W Panasonic microwave to generate steam, a single 2-minute treatment of the 3M 1860 and 3M 1870 N95 respirator models did not result in any gross physical damage.6 No microwave sparking from the metallic noseband material was observed in any studies.
Microwave-generated steam as a means of FFR decontamination is advantageous because of short exposure times and uses inexpensive, readily available equipment.60,61 However, if greater water reservoir volume is used to generate steam, more time is needed to produce the adequate amount of steam. Only 1 FFR can be processed at a time because of the small interior space of commercially available microwaves.
Ethylene oxide is a sterilant gas with bactericidal, sporicidal, and viricidal activity commonly used in health care for medical equipment and supplies.62 The broad material compatibility of ethylene oxide makes it compelling when considering N95 respirator decontamination, but risks are present involving safety and handling of a flammable and hazardous gas and residue control following sterilization.47 Furthermore, these processing systems may not be present at or close to many health care centers.
Seven studies evaluated ethylene oxide use on 15 different FFR models (Table 5). In 2 studies, the Steris AMSCO Eagle 3017 ethylene oxide sterilizer was used, and in 3 studies the 3M Steri-Vac (4XL and 5XL) gas sterilizer/aerator was examined. The systems operated at 54 °C to 55 °C with chamber volumes of 0.13 m3 to 0.14 m3. The ethylene oxide concentration ranged from 725 mg/L to 883 mg/L and treatment timed varied among 1 hour,8,54,55 3 hours,47 and 12 hours.29 Among these studies, 1 reported infectious outcomes for N95 respirators contaminated with vesicular stomatitis virus. Ethylene oxide effectively reduced vesicular stomatitis virus to undetectable limits after 1 hour of treatment.27
In the 6 studies assessing the effect of ethylene oxide on FFR filter performance, no significant changes were observed.8,27,29,54,55,63 While 1 to 3 cycles of processing caused no damage or change in physical appearance to the facepieces of the respirators, 725 mg/L to 883 mg/L of ethylene oxide concentrations caused the P100 (make/model unspecified) respirator straps to darken after 1 hour of treatment followed by 4 hours of aeration using the 3M Steri-Vac 4XL and 5XL ethylene oxide devices.54
The ability of high-volume reprocessing of FFRs with ethylene oxide for FFR decontamination has not been published in the peer reviewed literature. Ethylene oxide reprocessing of FFRs is limited by a lack of studies showing the safety and efficacy of this approach, the need for long aeration periods (>4 to 12 hours) to reliably remove hazardous ethylene oxide gas from FFR devices,8,27,29,47,54,55,63 and the need for expensive equipment.64 Ethylene oxide is a flammable, hazardous gas with known toxic by-products.47,65 The CDC recommends against the use of ethylene oxide for personal protective equipment reprocessing because of the potential for harm to the wearer. Detailed results on toxicity are available in eAppendix 4 in the Supplement.
The COVID-19 pandemic resulted in a critical shortage of FFRs that has persisted for nearly a year. Inadequate FFR supply has resulted in widespread reprocessing of these devices. N95 respirators are designed for single use, and there are few studies of the effectiveness of reprocessing them for repeated use. The CDC recommends FFR reprocessing only in extreme circumstances, which, unfortunately, have existed for the entire duration of the COVID-19 pandemic. Out of necessity, many clinicians and institutions have adopted protocols for FFR reprocessing, but many of these have not been studied. This review summarizes the available evidence regarding FFR reprocessing. Despite a modest body of available literature, only a few studies examined the effectiveness of FFR reprocessing on SARS-CoV-2 contamination, with the remaining studies evaluating the effect of decontamination procedures on other viruses and bacteria.
Ultraviolet germicidal irradiation, VHP, MHI, and MGS processing effectively sterilize FFRs and retain filtration performance (Box). Ultraviolet germicidal irradiation and VHP are advantageous because they have a low risk of damaging mask components. Physical degradation and changes to the mask may occur, depending on the FFR make and model and decontamination method,22,66 that may affect fit and decrease filter performance.21,63,67 When clinicians use a reprocessed N95 respirator, they should inspect the entire mask for visual damage and the elastic function of the straps, ensuring a proper seal and facial fit before using the mask (Box). They should ensure that whatever process is used to reprocess the mask has been tested to confirm that adequate filtration efficiency is maintained.30,68 Because FFR devices are made of a variety of materials and constructed in different ways, clinicians should establish that the reprocessing system in use at their facility has been tested for the specific make and model of the N95 respirator. The CDC recommends that the manufacturer be consulted about the effect of the chosen decontamination process method for the specific respirator model before reprocessing is attempted in the clinical setting. If the manufacturer cannot provide this information, the methods used to test the efficacy of reprocessing systems reported in the literature summarized in this review should be implemented by local facilities to ensure the safe operation of reprocessed FFR devices.
Ultraviolet germicidal irradiation (UVGI) is cost-effective and should be considered when equipment (UV lamps or bulbs, biosafety cabinets/laboratory hoods, or whole-room decontamination systems) are available. Vaporized hydrogen peroxide (VHP) is a comparable method and its equipment is often available because it is used for sterilizing other medical equipment (but cost is higher).
Filtration performance of FFRs is maintained with up to 9 to 10 cycles of whole-room UVGI decontamination (3 studies). The efficacy of pathogen decontamination has been demonstrated in 1 study.
Consider UVGI box reprocessing, as it appears to damage FFRs the least. If UVGI is unavailable, moist-heat incubation (MHI) using preheated containers/chambers and microwave-generated steam (MGS) are comparable options. With MHI, the Caron incubator should be avoided because of the greatest risk of mask damage. With MGS, efficacy is dependent on volume of water used in the reservoir to produce adequate amount of steam (50-100 mL).
No reduction in FFR filtration performance with ethylene oxide has been reported. However, ethylene oxide is a human carcinogen and its chemical residue is an unresolved concern; this method has not been proven safe.
Systematic evaluation of reprocessed FFRs should occur after decontamination methods are implemented. The FFR should be visually inspected for damage, filtration performance and facial fit assessed, and a proper seal ensured. Reprocessed FFRs should not be used for aerosol-generating procedures, such as intubation or bronchoscopy.
Decontaminated FFRs should be examined for visible signs of deterioration/damage such as changes in texture (softness, pliability, coarseness) and separation of inner foam nose cushion from FFR body.
The integrity of FFRs should be assessed (elastic function and breathing resistance) and qualitative fit testing performed. Filtering facepiece respirators donned more than 5 times should undergo qualitative FFR fit performance evaluation. Fit testing should be performed for new FFR models or change in decontamination protocol.
A user seal check should be performed by the wearer to ensure an adequate seal is achieved. If air leaks around the nose, the user should readjust the nose piece and straps. If air leaks between the facial seal, the FFR should be discarded.
Reprocessed FFRs should be discarded when soiled or damaged, if the mask creates more difficulty breathing through it, or if there is failure to achieve a proper fit/seal during user seal check. The Centers for Disease Control and Prevention limits donnings to 5, then disposing of the reused FFRs, unless the manufacturer states otherwise.
Normal operations should resume when supplies meet projected FFR demand.
Ultraviolet germicidal irradiation may be one of the most practical methods for reprocessing N95 respirators. Commercially available UV light boxes and laboratory hoods are currently used in health care systems and are readily accessible.31,32 Many health care systems have adopted whole-room UVGI decontamination systems. To date, there are limited peer-reviewed reports of this method’s efficacy.24,27,30 Closed box systems are simple to construct and may provide more reliable and reproducible UVGI dosage than whole-room decontamination systems.69 When using UVGI for FFR decontamination, the FFR material type, design, and location of contamination on the mask should be considered to determine the optimal means for exposing the mask to UV light.14,15 Ultraviolet light primarily decontaminates surfaces and is more effective when directed at inoculum dispersed over larger surface areas as opposed to droplets, in which clumping or stacking of organisms may occur.18 Hydrophilic and irregular surfaces, such as occur in Gerson 1730, Sperian HC-NB095, Dentec Safety AD2N95A, 3M 1860S, and Moldex-Metric 1517 N95 respirators, may not be amenable to uniform light exposure of critical surfaces, resulting in inadequate decontamination of FFR devices by UVGI.14,15,20,27
Ultraviolet doses of 1 to 2 J/cm2 effectively sterilize N95 respirators for most pathogens without causing degradation of mask components.6,20-22 However, in the 3 studies specifically assessing SARS-CoV-2, higher UV doses were necessary or used,13,14 so higher doses of UV energy might be needed for reprocessing FFRs during the COVID-19 pandemic. When reprocessing FFR using a UV box, both the interior and exterior mask surfaces must be exposed to the UV light to address potential contamination from either the external environment or the wearer. Mask material should be directly exposed to UV light and the amount of UV light delivered measured using a radiometer or colorimetric indicator. Because UV decontamination depends on exposure of the mask material to UV light energy, N95 respirators should not be stacked during the process.
Recently, the US Food and Drug Administration approved the use of VHP decontamination that is performed in either sterilization/aeration chambers or by mobile generators placed in large room spaces.70 Vaporized hydrogen peroxide effectively decontaminated FFRs impregnated with virus including SARS-CoV-2, bacteria, and spores.8,13,16,27,41 Filtering facepiece respirator mask integrity was maintained for up to 20 to 30 cycles of VHP open-room processing.41,43 Physical degradation and reduction in N95 respirator filter performance was observed using the STERRAD VHP system that uses small chamber sizes and higher fractions of H2O2,8,27,30 and early fit test failure was observed with extended clinical FFR use and decontamination with the Steris V-PRO maX VHP system.40 The use of self-seal pouches protects against VHP absorption of cellulose-based material when using small chambers. Throughput challenges using the Bioquell Clarus C hydrogen peroxide vapor generator and, in more recent studies, using VHP chamber systems have not been reported.13,30,39 Open-room sterilization and aeration may limit cycle abortion and FFR damage while maximizing the volume of N95 respirators that can be decontaminated during a single cycle. Low levels of residual oxidant may be present on FFRs after treatment, so residual hydrogen peroxide levels should be assessed41,47 and/or confirmed with the manufacturer to be below the established residue limit after treatment. Although this technology may support large-volume FFR decontamination efforts, scalability is limited by long treatment and aeration times and expense.
Moist-heat incubation and MGS effectively kill microorganisms by heating them. When generating moist heat using a water-filled container in an oven, preheating allows the liquid to reach the desired temperature prior to decontamination, whereas incubator and heated cabinets can be set to the desired temperature and relative humidity. These modalities distribute heat evenly across the surface being sterilized, facilitating the use of lower temperatures than otherwise needed for sterile processing and reducing the risk of damaging polymer fibers in N95 respirators. Sterilization and retention of FFR filtration and integrity are optimized by MHI performed with a relative humidity of 80% to 100% at 60 °C to 85 °C for 15 to 30 minutes or by MGS in an 1100-W to 1800-W device (the power range of commercially available microwave devices) with treatment times ranging from 40 seconds to 3 minutes.6,50,57,58 These techniques are useful for small-scale decontamination because the equipment is readily available and the treatment times are short. However, preheating time as long as 3 hours is required for MHI, and physical damage to the N95 respirator can occur from both techniques.8,23
Ethylene oxide was a common means for processing equipment not amenable to steam sterilization. It fell out of favor because of risks with the handling of flammable and hazardous gases and residual potential toxic by-products.47,65 Nevertheless, it can be used to reprocess N95 respirators. Studies show that ethylene oxide effectively sterilizes bacteria and viruses from FFRs and surfaces,27,62 and filter performance is preserved after 1 to 3 cycles of ethylene oxide treatment.8,27,29,63 Lengthy aeration cycles lasting more than 4 hours are needed to vent toxic ethylene oxide, limiting how many masks can be processed in a timely manner. The CDC’s guidance recommends against the use of ethylene oxide due to the risk of inhalation by the wearer, as its carcinogenic and potentially teratogenic toxicity are unresolved concerns.
N95 respirators are highly effective at reducing the transmission of viral infection.2 However, when the demand for these devices is unusually high, as has occurred during the COVID-19 pandemic, there may be a need to reuse these devices that were originally intended for single use. Fortunately, research performed to date has shown that as a general proposition, N95 respirator decontamination processes are feasible. Ultraviolet germicidal irradiation, VHP, MHI, and MGS processing can sterilize N95 respirators while retaining filtration performance and fit. Ultraviolet germicidal irradiation and VHP appear to cause the least physical damage to the respirator components. Of these modalities, UVGI might be the most feasible approach for FFR reprocessing because UV bulbs are ubiquitous in the medical environment, the required equipment is readily available, and the process is safe. Vaporized hydrogen peroxide is a suitable alternative for high-throughput processing but requires relatively expensive equipment and may be difficult to implement. Although decontamination can be achieved with these techniques, FFRs were not designed for reuse and should be processed as few times as necessary until unused FFR devices can be obtained.
This review has several limitations. First, the included studies were experimental, laboratory-based studies, with the exception of 2 studies that evaluated respirator performance in clinical use.13,40 Second, only 5 studies demonstrated adequate reduction of infectivity for SARS-CoV-2. Most of the studies examined other viral and bacterial pathogens. Third, there are more than 530 types of NIOSH-approved N95 respirators, and findings may not generalize to all of them. The lack of evidence pertaining to decontamination for specific types of N95 respirators suggests that individual institutions should test their specific FFRs to ensure pathogen removal and retained mask filtration efficiency and facial fit. A hospital’s health and safety and infection control office should be consulted.
The COVID-19 pandemic has been characterized by persistent shortages of personal protective equipment, especially N95 respirators. As the supply of these masks has remained limited throughout the duration of the pandemic, it can be assumed that in future pandemics, reprocessing of FFRs will be necessary. Health care institutions should have plans in place to implement one of the FFR reprocessing systems described in this review. They should be prepared to test the effectiveness of the chosen reprocessing method for the specific pathogens of concern and masks used at their institution before implementing large-scale reprocessing programs. Given the inevitability of the need for reprocessing in emergency systems, FFR manufacturers should provide guidance for how to best reprocess their FFR devices when emergency conditions exist.
Ultraviolet germicidal irradiation, vaporized hydrogen peroxide, moist heat, and microwave-generated steam processing effectively sterilized FFRs and retained filtration performance. Ultraviolet irradiation and vaporized hydrogen peroxide damaged masks the least. More research is needed on decontamination effectiveness for SARS-CoV-2 because few studies specifically examined this pathogen.
Corresponding Author: Max A. Schumm, MD, Department of Surgery, UCLA David Geffen School of Medicine, 10833 Le Conte Ave, 72-227 CHS, Los Angeles, CA 90095 (email@example.com).
Accepted for Publication: February 11, 2021.
Published Online: March 3, 2021. doi:10.1001/jama.2021.2531
Author Contributions: Dr Schumm had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Schumm, Hadaya, Mody, Myers.
Acquisition, analysis, or interpretation of data: Schumm, Hadaya, Maggard-Gibbons.
Drafting of the manuscript: Schumm, Hadaya.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Schumm, Hadaya.
Conflict of Interest Disclosures: None reported.