Included studies

None of the 15 included studies assessed surgical masks. All studies assessed FFRs, including N95 and P100. All included studies were nonhuman subject research. Of the 15 studies, 4 studies (27%) assessed disinfection of bacteria, ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17,21,Reference Schwartz, Stiegel and Greeson22} 7 studies (47%) assessed disinfection of virus, ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15,Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18,Reference Fisher, Williams and Shaffer24} 9 studies (60%) assessed postdecontamination filtration efficiency, ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15,Reference Lindsley, Martin and Thewlis16,Reference Viscusi, King and Shaffer19–Reference Schwartz, Stiegel and Greeson22,Reference Fisher, Williams and Shaffer24} and 12 studies (80%) assessed physical structure degradation. ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman, Eimer and Shaffer11–Reference Viscusi, Bergman and Novak14,Reference Lindsley, Martin and Thewlis16,Reference Viscusi, King and Shaffer19–Reference Fisher, Williams and Shaffer24}

We identified 14 decontamination methods. Data regarding ultraviolet germicidal irradiation (UVGI) (9 studies) ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman, Eimer and Shaffer11–Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} are described in Supplementary Table 1 (online). Data regarding moist heat (5 studies) ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} are described in Supplementary Table 2 (online). Data regarding microwave-generated steam (MGS) (4 studies) ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} and hydrogen peroxide vapor (HPV) (4 studies) ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,21–Reference Kenney, Chan and Kortright23} are described in Supplementary Table 3 (online). Data regarding microwave steam bags (1 study), ^{Reference Fisher, Williams and Shaffer24} bleach (5 studies), ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} steam treatment, (3studies), ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} dry heat (3 studies), ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} ethanol or isopropyl alcohol (3 studies) ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} , ethylene oxide (EtO) (3 studies), ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} hydrogen peroxide gas plasma (HPGP) (2 studies), ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Viscusi, King and Shaffer19} liquid hydrogen peroxide (LHP) (2 studies), ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} microwave irradiation (1 study), ^{Reference Viscusi, King and Shaffer19} and soap and water (1 study) ^{Reference Viscusi, King and Shaffer19} are described in Supplementary Table 4 (online).

Ultraviolet germicidal irradiation

Overall, 9 studies assessed the performance of UVGI decontamination method. ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman, Eimer and Shaffer11–Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} All studies evaluated ultraviolet light-C (UV-C) with a primary wavelength of 254 nm. However, a great variety of equipment and delivery techniques were employed: a laminar-flow cabinet with UV-C light, ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} a UV-C lamp, ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17} and a chamber with a UV-C bulb. ^{Reference Lindsley, Martin and Thewlis16,Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} Diversity across studies included the intensity of UV-C (W/cm²), the dose of UV-C (J/cm²), the distance between the source of UV-C and the FFRs, the exposure surface of the FFRs, the total exposure time, and the number of cycles. The exposure time per cycle varied from 1 minute, ^{Reference Lindsley, Martin and Thewlis16–Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} 15 minutes, ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} 20 minutes, ^{Reference Lin, Tang, Hung, Hua and Lai17} 30 minutes ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Viscusi, Bergman and Novak14} to 45 minutes. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12} The number of cycles varied from 1 ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Viscusi, Bergman and Novak14–Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} to 3 cycles. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13}

UVGI was effective for influenza virus inactivation, including H5N1 (2 studies) ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} and H1N1 (3 studies), ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13,Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} and Bacillus subtilis spore inactivation. ^{Reference Lin, Tang, Hung, Hua and Lai17} Post-decontamination filtration efficiency was unchanged in 4 studies. ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15,Reference Lindsley, Martin and Thewlis16} The physical structure was unchanged in 3 studies, ^{Reference Viscusi, Bergman, Eimer and Shaffer11–Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13} but physical strength partially lost with high doses of UV-C at 120 J/cm² and 950 J/cm² and the head strap strength lost at 590 J/cm². ^{Reference Lindsley, Martin and Thewlis16} The optimal UV-C dose should be <2 J/cm². Laminar flow cabinet was suggested for 3M 1860, 3M 1870, Kimberly Clark PFR 95-270. ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} UV-C chamber was suggested for 3M 1860, 3M 9210, Gerson 1730, Kimberly Clark 46727. ^{Reference Lindsley, Martin and Thewlis16,Reference Mills, Harnish, Lawrence, Sandoval-Powers and Heimbuch18} UV-C lamp was effective, but the FFRs model was unspecified. ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17}

Moist heat

In 5 studies, the performance of moist heat decontamination method was assessed, ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} and 2 types of equipment were used. In 3 studies, N95 FFRs were decontaminated with a laboratory incubator for a 30-minute incubation at 60°C. The FFRs were air dried after each incubation: overnight after the first incubation and for 30 minutes after the second and the third incubations. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12–Reference Viscusi, Bergman and Novak14} In 2 studies, a 6-L sealable container filled with 1 L tap water was warmed in a 65°C oven for a minimum of 3 hours. Then the FFRs were placed on a rack to isolate the FFRs from the liquid, and the containers were sealed and returned to the oven. ^{Reference Heimbuch, Wallace and Kinney8,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} The exposure time per cycle ranged from 15 minutes ^{Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13} to 20 minutes ^{Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} to 30 minutes. ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, Bergman and Novak14} The number of cycles ranged from 1 ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman and Novak14,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} to 3. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13}

Moist heat was effective for the H1N1 ^{Reference Heimbuch, Wallace and Kinney8} and H5N1 ^{Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} influenza virus inactivation when using a prewarmed sealable container. Viral inactivation was not achieved by a laboratory incubator. Bacterial disinfection was not assessed. The postdecontamination filtration efficiency was >97.5%. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} The physical structure was reported unchanged by 1 study ^{Reference Heimbuch, Wallace and Kinney8} but degradation was reported in some models by 3 studies. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12–Reference Viscusi, Bergman and Novak14} A 6-L prewarmed sealable container filled with 1 L tap water was suggested for models 3M 1860 and 3M 1870. ^{Reference Heimbuch, Wallace and Kinney8,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15}

Microwave-generated steam

In 4 studies, the performance of MGS decontamination was assessed. ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} FFRs were placed outer-side down on top of 2 side-by-side pipette tip boxes with 50 mL room-temperature tap water, in a 1,100W, 1,250W (2,450 MHz) microwave oven with a revolving glass carousel. The exposure time was 2 minutes at the maximum power setting. Then the FFRs were dried overnight on a laboratory benchtop. In 3 studies, the FFRs were decontaminated with 1 cycle, ^{Reference Heimbuch, Wallace and Kinney8,Reference Viscusi, Bergman and Novak14,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} and in 1 study FFRs were decontaminated with 3 cycles. ^{Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13} MGS inactivated >4-log reduction of the viable virus. ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13–Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} Bacterial disinfection was not assessed. The post-decontamination filtration efficiency was unchanged. ^{Reference Lore, Heimbuch, Brown, Wander and Hinrichs15} In 3 studies a slight separation of the inner-foam nose cushion was observed in some samples. ^{Reference Heimbuch, Wallace and Kinney8,Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13,Reference Viscusi, Bergman and Novak14} Although a minor physical structure degradation was reported, the FFRs had a 90%–100% fit-test passing rate. ^{Reference Bergman, Viscusi, Zhuang, Palmiero, Powell and Shaffer13} A 1,250W (2450 MHz) microwave oven with a revolving glass carousel was suggested for 3M 1860. ^{Reference Heimbuch, Wallace and Kinney8,Reference Lore, Heimbuch, Brown, Wander and Hinrichs15}

Hydrogen peroxide vapor

In 4 studies, the performance of HPV decontamination was assessed for 3, 5, 30, and 50 cycles (225, 405, and 480, and 1,440 minutes per cycle). ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,21–Reference Kenney, Chan and Kortright23} An HPV generator utilizing 30% or 35% hydrogen peroxide solution was placed in a room. The FFRs were placed on stainless-steel wire racks. The HPV run consisted of the following 5 stages: conditioning, pre-gassing, gassing, gassing dwell, and aeration. The processing room attained the 480+ parts per million (ppm) level of HPV with gassing times of 25 and 40 minutes and gassing dwell times of 15, 20, and 25 minutes (ie, the sterilization process). During the aeration stage, fresh air was introduced into the room to increase the rate of catalytic conversion of hydrogen peroxide into water and oxygen. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,21–Reference Kenney, Chan and Kortright23} In addition, 4 hours of aeration eliminated the toxicity of hydrogen peroxide. ^{Reference Schwartz, Stiegel and Greeson22} Viral disinfection was not assessed. HPV was effective for Geobacillus stearothermophilus spore inactivation. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,21,Reference Schwartz, Stiegel and Greeson22} Post-decontamination filtration efficiency ^{21,Reference Schwartz, Stiegel and Greeson22} and the physical structure were unchanged. ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Schwartz, Stiegel and Greeson22,Reference Kenney, Chan and Kortright23} Placement of the FFRs on stainless-steel wire racks in the room with an HPV generator was suggested for 3M 1860. ^{21,Reference Schwartz, Stiegel and Greeson22}

Microwave steam bags

One study assessed the performance of microwave steam bag decontamination. ^{Reference Fisher, Williams and Shaffer24} The FFRs were placed inside separate bags filled with 60 mL tap water. The bags were sealed, placed in a microwave oven, and irradiated on high power for 90 seconds. Bacteriophage MS2, a surrogate for a pathogenic virus, was thereby inactivated, but bacterial inactivation was not assessed. The postdecontamination filtration efficiency and the physical structure were unchanged. ^{Reference Fisher, Williams and Shaffer24}

Bleach

The performance of bleach decontamination was assessed in 5 studies. ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} The FFRs were submerged in a 0.6% aqueous solution of sodium hypochlorite for 1–3 cycles. The exposure time ranged from 10 to 30 minutes. ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} After treatment, they were hung on a laboratory pegboard and allowed to air dry overnight. Virus inactivation was not assessed. Bleach was effective for Bacillus subtillis spore inactivation. ^{Reference Lin, Tang, Hung, Hua and Lai17} The postdecontamination filtration efficiency was unchanged in 3 studies, ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} but it decreased in 1 study. ^{Reference Lin, Chen, Huang, Kuo, Lai and Lin20} The physical structure of N95 FFRs was degraded after the 30 minutes of decontamination. ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19}

Steam

The performance of steam decontamination was assessed in 3 studies. ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} Characteristics of these studies assessing steam are described in Supplementary Table 5 (online). The FFRs were sealed in an autoclave bag and treated in an autoclave at 121°C. The FFRs were air dried for 72 hours. Virus inactivation was not assessed. Steam was effective for Bacillus subtillis spore inactivation. ^{Reference Lin, Tang, Hung, Hua and Lai17} The filtration efficiency decreased in 2 studies, ^{Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} and the outer layer of the N95 FFRs was deformed, shrunken, and stiff. ^{Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20}

Dry heat

The performance of dry heat decontamination was assessed in 3 studies ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} using 2 types of equipment. One study used a hot-air oven; respirators were placed in a metal pan on racks of a laboratory oven and were turned over midway through the exposure period (60 minutes) for 1 cycle at 80°C and 160°C. ^{Reference Viscusi, King and Shaffer19} Two studies used an electric rice cooker at 149–164°C for 3 minutes for 1 cycle, ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} but virus inactivation was not assessed. Dry heat with an electric cooker was effective on the disinfection of Bacillus subtilis spores. ^{Reference Lin, Tang, Hung, Hua and Lai17} The postdecontamination filtration efficiency was unchanged, but the FFRs were melted at 160°C after 22 minutes of decontamination. ^{Reference Viscusi, King and Shaffer19}

Ethanol or isopropyl alcohol

The performance of ethanol or isopropyl alcohol decontamination was assessed in 3 studies. ^{Reference Lin, Tang, Hung, Hua and Lai17,Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} Ethanol with various concentrations and volumes was added to the center of the surface of the N95 FFRs. The FFRs were then dried in a petri dish placed in a biosafety cabinet for 10 minutes, followed by another 10 minutes of submersion in 100% isopropanol solution. Virus inactivation was not assessed. Ethanol was effective in the disinfection of Bacillus subtilis spores. ^{Reference Lin, Tang, Hung, Hua and Lai17} The postdecontamination filtration efficiency decreased, ^{Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20} and the physical structure was unchanged. ^{Reference Viscusi, King and Shaffer19,Reference Lin, Chen, Huang, Kuo, Lai and Lin20}

Other methods

Other decontamination methods assessed neither viral nor bacterial inactivation. Post-decontamination filtration efficiency remained unchanged for EtO, ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} HPGP, ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Viscusi, King and Shaffer19} LHP, ^{Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} microwave irradiation decontamination. ^{Reference Viscusi, King and Shaffer19} In contrast, soap and water decreased filtration efficiency. ^{Reference Viscusi, King and Shaffer19} Physical degradation was shown in the ethylene oxide, HPGP, LHP, microwave irradiation methods ^{Reference Viscusi, Bergman, Eimer and Shaffer11,Reference Bergman, Viscusi, Heimbuch, Wander, Sambol and Shaffer12,Reference Viscusi, King and Shaffer19} but the physical structure was unchanged for soap and water. ^{Reference Viscusi, King and Shaffer19}