Mask decontamination methods (model N95) for respiratory protection: a rapid review

Based on these recommended approaches [25], we developed a specific protocol for this study (Supplementary material 1). We planned to include only primary research studies that evaluated methods for the decontamination of N95 masks for reuse and whose outcome was the effectiveness, safety, maintenance of protection, or filtering characteristics of the evaluated decontamination method were included. Therefore, we could answer the question: “How effective and safe are decontamination methods for respiratory protection masks model N95/PFF2 against respiratory viruses?”

We followed the limit main database searching recommended by Cochrane Rapid Reviews Methods Group [27]. Therefore, searches were conducted on MEDLINE, Cochrane Library, and EMBASE databases on September 25, 2020. Search terms were related to decontamination (e.g., “Sterilization,” “Disinfection,” and “Decontamination”), reuse (e.g., “Equipment Reuse” and “Reuse”), device failure (“Equipment Failure”), and masks (e.g., “N95” and “filtering facepiece respirators”).

The search strategy was developed in two stages. In the first stage, an experienced researcher (LFP) structured the strategies with the collaboration of different specialists (AIQC, ATGG, JV, MCCF, AMMP, JOB, and SMVLO). In the second stage, a researcher (AJG) who is trained by the Cochrane Systematic Reviews Group evaluated and validated the search strategies. The review did not have a date or language restrictions. The complete search strategies can be found in Table 1 of Supplementary material 2; therefore, they are valid, reliable, and reproducible.

In addition to searching the official databases, the reference lists of all studies selected for full-text reading, as well as the review articles identified in official searches, were scrutinized to identify possible eligible studies.

Two reviewers (LFP and AIQC) independently screened for the title, abstracts, and full text, the reviewers used the Rayyan systematic review application in blind mode [28]. Disagreements during the selection process were resolved by discussion with a third review (SMLVO). Mendeley citation management software was used for the automatic removal of duplicate articles.

Two reviewers (LFP and AIQC) independently used a pre-specified data extraction sheet form, in duplicates, designed to obtain the specific data required for this review. The data extracted from the primary studies were data related to the author, year, study objective, intervention, comparator, commercial mask model, target microorganism, results, and conclusions of the study authors, limitations, and detailed description of the decontamination process making its reproducibility in other scenarios possible. Moreover, authors of the included primary studies were also contacted to provide data that was not available in the manuscripts.

The characteristics of each study (cycles, temperatures, protocols, densities, exposure time, technology used, and results) are presented in Table 3 of Supplementary material 4 [29‐76]. The differences before and after decontamination are shown in Table 4 of Supplementary Material 5 for outcomes: filter aerosol penetration, filter airflow resistance, and filtration efficiency. Results of the studies were summarized based on the decontamination method and results for the two following issues [53]:

The complete search strategies can be found in Table 1 of Supplementary material 2. Initial searches retrieved 552 articles (MEDLINE: 381, Cochrane: 52, and EMBASE: 119). The reference lists of all studies selected for full-text reading, as well as the review articles identified in official searches, were scrutinized to identify possible eligible studies. Authors of the included primary studies were also contacted. These manual searches detected 11 additional publications that were added to the total recovered. Mendeley citation management software was used for the automatic removal of duplicate articles, leaving 301 studies remaining.

Two reviewers (LFP and AIQC) independently screened the 301 studies using the Rayyan systematic review application to screen abstracts and titles [28]. Of these, 240 were excluded for not meeting the inclusion criteria. The full texts of 61 studies were screened by two reviewers (LFP and AIQC), and thirteen additional studies were excluded. Excluded full-text studies and the reasons for exclusion are listed in Table 2 of Supplementary material 3 [79‐91]. At the end, 48 studies were selected for the full-review process (Fig. 1).

Fifteen methods were assessed in the 48 papers: hydrogen peroxide, ultraviolet irradiation, ethylene oxide, dry heat, moist heat/pasteurization, ethanol, isopropanol solution, microwaving, sodium hypochlorite (NaClO), autoclaving, electric rice cooker, cleaning wipes, bar soap, and water, multi-purpose high-level disinfection cabinet (Altapure, Mequon, WI), and chlorine dioxide (ClO₂) [29‐76]. Each method will be briefly analyzed below:

Hydrogen peroxide was evaluated in its liquid, plasma, and gas/vapor forms by nineteen laboratory studies [29‐47]. The effect of hydrogen peroxide on the filtering capacity varied according to the method used. The average penetration was not significantly changed when the masks were submerged in 3% or 6% hydrogen peroxide liquid or treated with vaporized hydrogen peroxide (STERRAD®) for one cycle [29]. However, it resulted in mean penetration levels > 5% after 3-Cycles [32]. The integrity and filtering capacity of the mask were preserved when hydrogen peroxide was used as steam [32, 34, 92].

Additionally, studies have shown hydrogen peroxide led to changes in the masks’ metallic nasal clips [29, 30, 32]. As for the ability to eliminate microorganisms, SteraMist™ Binary Ionization Technology® (BIT™) was effective against influenza A virus subtype H1N1 [33] and Geobacillus stearothermophilus spores [43] and STERRAD 100NX sterilization system eliminated SARS-CoV-2, Staphylococcus aureus, and Acinetobacter baumannii [44].

On the other hand, vaporized hydrogen peroxide (VHP) was effective in eliminating SARS-CoV-2 [34], Geobacillus stearothermophilus spores [35], Porcine respiratory coronavirus (PRCV) [39], Escherichia coli, Mycobacterium smegmatis, and spores of Bacillus stearothermophilus [45] and 3 aerosolized bacteriophages: T1, T7, and Pseudomonas phage phi-6 [47].

The effect of ultraviolet germicidal irradiation on N95 respirator masks was evaluated by 21 studies [29‐32, 34, 39, 42, 48‐61], and there was a difference among studies in relation to UVGI doses and application time periods. In general, UVGI did not affect the integrity and ability of the masks to filter aerosols or adapt to the face, nor did it leave a smell, irritating/toxic residues, or create important changes in appearance even when multiple cycles were performed [29‐32, 34, 42, 49, 51‐53, 55, 57, 58]. However, different commercial brands of N95 models resisted differently in terms of performance penetration after multiple cycles and doses applied [55]. The most efficient N95 brand in terms of long-standing penetration performance was 3M 9210, but the mean penetration values for all brands were 5% or less both before and after exposure [55]. The most efficient N95 brand in terms of long-standing flow resistance performance was 3M 9210, but the mean flow resistance values were less than 1% of the initial value for all brands [55].

Ultraviolet germicidal irradiation was effective against MS2 coliphage (ATCC 15597-B1) [48], the influenza virus H5N1 [53] SARS-CoV-2 [34, 60, 61], bacteriophages MS2 [49, 54], influenza virus H1N1 [50, 56] and porcine respiratory coronavirus (PRCV) [39]. However, at the same time, one study reported that even after 20 min of irradiation with 365 nm UVA the relative survival of Bacillus subtilis spores remained above 20% [57] and another one highlighted the UV-C technologies tested did not meet pre-established criteria for decontamination against Methicillin-resistant Staphylococcus aureus (MRSA), Bacteriophage Phi6, and Bacteriophage MS2 [59].

Evaluated by four studies, the effectiveness of ethylene oxide (EtO) depended on the type of sterilization equipment used, whether there was a hot cycle, and exposure to EtO [29‐32]. The process did not affect the filtration, resistance, odor, or appearance of the masks. The main limitations of the method were the processing time and the presence of toxic residues. It is also important to note that none of the studies tested the effectiveness of EtO treatment against microorganisms. Thus, there is no evidence that EtO can eliminate any microorganism from N95 masks.

The use of dry heat was evaluated by eleven different experimental studies [29, 30, 34, 39, 58, 59, 62‐66]. Temperatures between 70 and 85°C did not affect the structural characteristics of the masks under various humidity conditions (≤100% RH) [34, 58]. Also, filtering efficiency remained acceptable (≥95%) up to 50 cycles at 85°C and 30% of RH [58]. When an oven was used at 70°C, filtration performance was maintained if only one cycle was performed [34]. However, there were no detectable changes in aerosol filtration efficiency, even after three reprocessing cycles when the masks were subjected to the same temperature of 70°C when laboratory MINI/6 incubator (Genlab Ltd.) was used to provide dry heat [63].

Dry heat treatment (70°C for 60 min) was effective against SARS-CoV-2 [34, 63]; however, dry heat (70°C for 30 min) had limited effectiveness against bacteriophages MS2 and Phi6 versus methicillin-resistant Staphylococcus [59]. Also, dry heat at 100°C for 15 min did not eliminate Methicillin-resistant Staphylococcus aureus (MRSA) and Nonenveloped single-stranded RNA virus bacteriophage MS2 [64].

Nine studies [32, 50‐53, 58, 63, 67, 68] evaluated the effect of moist heat between 60 and 100°C. The method did not alter mask fit, odor, or comfort [32, 50, 51, 53, 58]. In one study [58], filtration efficiency presented a significant drop after cycle 5 when stacked on top of a beaker with boiling water inside (around 15 cm above the water), although filtration efficiency was not affected when masks were subjected to 5 cycles with a lower temperature—at 85°C + 60–85% humidity [68]. Moist heat (65±5°C for 3h) was effective in eliminating H1N1 [50] and H5N1 [53] viruses; in addition, when masks were steamed in boiling water, the treatment was effective against the Avian infectious bronchitis virus H120 without changing [67]. Another study also has shown that a single-heat treatment for thermal disinfection in cycles of 60 min at 70°C at 50% eliminated SARS-CoV-2 and Escherichia coli [63].

Different methods of decontamination with ethanol were tested: spray [34], immersion [58, 69, 70], and pipette drips [57]. Results were divergent between methods. The filtration efficiency of masks was degraded to unacceptable levels when they were immersed in alcohol [58, 69, 70]. Mask filtration performance was not significantly reduced after single ethanol sprays which were also effective in eliminating SARS-CoV-2 [34]. Subsequent rounds of spraying caused a significant decrease in efficacy in filtration performance [34]. Nonetheless, pipette drips were not effective in eliminating Bacillus subtilis spores [57].

The filtering capacity of N95 masks was changed and the particle penetration through N95 masks exceeded 5% after they had been submerged in isopropanol solution [29, 69]. Although effects on microorganisms were not evaluated, this method could be further studied in order to check these effects.

Eleven studies tested the use of microwave ovens in the disinfection of N95 masks [29, 30, 32, 49‐53, 62, 71, 72]. The type of commercial furnace, maximum temperature, and time protocols varied between the studies (Table 3, Supplementary material 4). When masks were placed directly on the rotating plate of the microwave without protection, two commercial models of the tested masks melted [30]. When the masks were placed in containers with water [32, 50, 51, 53, 72] or in steam bags specifically marketed for microwave ovens no residual odor was observed [49]. In addition, there were no structural changes affecting adjustment on the face, filtration capacity, or resistance to airflow and none of the metal components melted or combusted. Microwaving the masks was effective in eliminating H5N1 [30] and H1N1 influenza viruses [51], bacteriophage MS2 [49, 72], and Staphylococcus aureus [62].

Ten studies [29‐32, 42, 48, 57, 58, 69, 71] evaluated the use of hypochlorite at different concentrations and application methods (Table 3, Supplementary material 4). The maintenance of mask integrity and filtering capacity varied among studies. For instance, one round of disinfection drastically degraded filtration efficiency to unacceptable levels when samples were left to air dry and off-gas completely, hanging [58], or submersed in 0.5% sodium hypochlorite for 10 min [69]. Application of sodium hypochlorite discolored the metallic components of the masks which, unfortunately, caused a characteristic smell of bleach [30].

One of the study treatments caused the release of low levels of hydrochloric gas [31]. Hypochlorite concentrations of 0.006%, 0.06%, and of 0.6% were not effective against Coliphage MS2 [71]. On the other hand, when higher sodium hypochlorite doses (>8.25 mg/liter) were used for the dilute solutions containing 2.75 to 5.50 mg/liter during a 10-min decontamination period MS2 coliphage was inactivated [48]. When 5.4%, 2.7%, or 0.54% NaOCl was used, the method was effective in eliminating Bacillus subtilis spores [57].

Considered an accessible method, as it is equipment present in all hospital environments, autoclave decontamination was evaluated by 7 studies [29, 57, 69, 70, 73‐75]. Autoclave disinfection was effective in eliminating Bacillus subtilis spores [57]; however, the integrity of the mask was altered [29, 69, 75]. Two studies pointed that even though masks were able to retain their structural integrity and efficacy, no results were provided regarding the filtering capacity, and only the fittest was performed [73, 74]. In addition, a survey conducted in 2020 showed the degree of integrity of the mask can be preserved depending on the model tested (Table 3, Supplementary material 4) [70].

The method showed 99–100% biocide efficacy against Bacillus subtilis [57] spores after using dry heat for 3 min (149–164°C, without adding water). Also, the treatment for 13–15 min, including 8–10 min of heating and 5 min of steam, resulted in a greater than 5 log₁₀ reduction in bacteriophage MS2 and methicillin-resistant S. aureus [64]. However, the method visibly changes the mask’s integrity [69].

The effectiveness of commercial wipes containing 0.9% hypochlorite, benzalkonium chloride, or no active antimicrobial ingredients was evaluated in masks contaminated with Staphylococcus aureus and mucin [76]. The three mask models, 3M-1860S, 3M-1870, and Kimberly-Clark-46727-PFR, withstood handling and abrasion during the disinfection process. All were successfully decontaminated against atypically high microbe levels by wipes containing antimicrobial agents; however, inert wipes did not produce adequate decontamination.

Average penetration has markedly increased for N95 respirators after being submerged. Authors have hypothesized the soap could have removed the charge on the fibers similar to the effect observed with isopropanol solution exposure [29].

The treatment was effective against microorganisms, and the researchers reported no visible changes in the masks. However, the efficiency of the filtration has not been confirmed [59].

The method significantly changed the filtering efficiency of the tested masks which makes this method not worth using or testing in further studies [36].

As reported in the results, the search strategy was developed by an experienced researcher and validated by another systematic review researcher, who is trained by the Cochrane Systematic Reviews Group. However, it was not developed by a health information specialist. Considering that this was the review that found the largest number of studies that were in fact eligible and included compared to previous reviews which investigated different mask decontamination methods (model N95) [18, 19], we do not consider this a limitation. As a rapid review, another limitation concerns a series of methodological simplifications adopted, which may affect the findings and our interpretations. Eliminating the evaluation of the studies’ methodological qualities was among the simplifications and calls for caution in interpreting the results presented. The technical aspects of outcome assessment in the included studies were not taken into account in this review and is an important limitation when interpreting the presented results.

The strengths of this review, however, are the number of identified decontamination methods and its unfolding results which points to practices that may be adopted and further studied as they seem to present better results initially, the multiple approaches used to search for relevant studies, such as contact with various authors and manual search of article references, and the participation of a team of multi-disciplinary specialists in all stages of the project, which included professionals from the areas: nursing, dentistry, medicine, pharmacy, and physical therapist.

A previous systematic review that included 15 studies concluded that future studies were required in order to establish the efficacy and safety of N95 decontamination methods [18]. Other systematic reviews have reported that masks can be decontaminated with microwave irradiation and moderate-temperature heat (up to 90°C), in both moist and dry conditions [20] and a single cycle of vaporized H₂O₂ can be used as a chemical disinfectant to remove viral pathogens without degrading the masks [22]. Our review is broader than the cited ones, which shows that our search strategy is much more sensitive. However, there is not sufficient evidence concerning UVGI as a safe decontamination method [21] To our knowledge, this is the first rapid review to assess 15 different decontamination methods identified in 48 studies, providing an overview of all available methods.

Whereas the current evidence is insufficient to determine a safe and widely accessible method, even for countries with financial limitations, our review points to an important gap in the evidence base, despite recent research efforts. In addition, considering the possibility of new challenging pandemic scenarios, investigating decontamination methods for reuse of protection professional equipment has an environmental, social appeal, and economic aspect. Moreover, it is important to invest in new fabric technologies that are prepared to be reused.