We observed that methylene blue and light applied to FFR and MM completely inactivates SARS-CoV-2. Importantly, MBL, as a new proposed decontamination method, did not cause any significant changes in any of the filtration, breathability or fluid resistance properties investigated in this study after 5 decontamination cycles, and our results are consistent with filtration efficiency seen after VHP+O₃ treatment. The additional finding that MB-pretreated FFRs and MMs continue to inactivate coronavirus provides a potential opportunity for healthcare providers and other front-line workers to safely re-use their PPE, as well as supply active viral inactivation during wear. MB is available globally and can be applied using spray bottles. This study demonstrates viral inactivation using MB combined with 50,000 lux of broad-spectrum light (∼1/3 of illumination on a sunny day). Bright light treatment can be supplied by simple white light lamps sold in hardware stores. Furthermore, in a small preliminary experiment, even MB exposed to ambient light (700 lux) and low light (<100 lux) inactivated SARS-CoV-2. This finding needs to be further studied but suggests that MB can work using ambient light available in hospital and clinic settings. We did observe, however, that lower light intensities require a longer light exposure duration for efficacy.

Identifying methods to decontaminate PPE to protect healthcare personnel is paramount. A recent study demonstrated recovery of viable virus from FFR material inoculated with viral loads of SARS-CoV-2 and stored at 20°C with 40-50% relative humidity for 21 days. (12) These common conditions in health care settings demonstrate the importance of having an inexpensive, simple, and robust means of decontamination of PPE like MBL that can retain functionality and performance. In our study, PRCV was inoculated under the outer layer of MMs and FFRs to mimic the worst-case scenario of high-titer virus penetrating the outer layer of a mask. MBL robustly inactivated >5 logs of virus for both MMs and 2 of the FFRs, suggesting that both MB and light were able to sufficiently penetrate the outer layer of the PPE. The exceptions were RM (2.7 log or 99.8% reduction) and CM (0.4 log or 60.5% reduction). RM is a half-sphere with a more rigid shell which may have blocked some of the MB or light from penetrating to the inner layer. Fisher et al. saw a similar finding with UVC light penetration, which was impacted the further into the filter material the light passed. (13)

Consistent with previous studies (14), the DH and VHP+O₃ treatments did not alter the filtration efficiency of masks and respirators tested in this study. MB and the DH decontamination can now be added to the rare category of decontamination methods that do not violate mask integrity for the models tested (including fit), which thus far only included the gold standard VHP (3,9,14-16-). Although the FDA approved a few VHP systems for PPE decontamination, they are not easily obtainable in low-resource settings. Various studies have also investigated the impact of heat on the filtration efficiency and physical deformation of FFR or filter media and demonstrated that higher temperatures could negatively affect integrity (13,17,18).

Dry heat yielded inconsistent viral inactivation in our study possibly owing to evidence that humidity plays a role in heat inactivation (19). The finding that DH on CM materials yielded better viral inactivation may be due to the ability of the cotton materials to rise to inactivation-level temperatures faster than the synthetic FFR and MM materials, although we did not test the material surface temperature during the heating to prove this. In one study, improved inactivation was demonstrated with moist heat (100% relative humidity) reducing initial SARS-CoV-2 titers almost 5-logs at 75°C for 30 minutes (19). Our rationale for testing 75°C DH without humidity related to testing the simplest method which could be replicated in any location where standard ovens exist without the complexity of regulating or delivering a certain amount of humidity.

We decided to compare a CM in this study to provide first definitive analysis in the same study format used for respirators and medical-grade masks. The tested CM performed surprisingly well in some of the integrity tests. This is likely due to the double cotton layer combined with two layers of polypropylene spunbond fabric. Not all CMs in use around the world contain these multiple layers, especially non-woven layers. The heterogeneity of materials and design, and lack of standardization of CM used for source control make generalized conclusions difficult. When testing MBL virucidal activity on CM, we observed inconsistent results when virus was inoculated under the cloth layer. This could be due to absorption of MB into the cotton layer. Alternatively, CM material color or finishing dye chemicals potentially impeded MB efficacy through light absorption. These hypotheses should be tested using alternative CM fabrics, colors, and models.

Our studies on MBL, DH, and VHP+O₃ decontamination were constructed to parallel other studies investigating the impact of heat and VHP+O₃ on the filtration efficiency and physical deformation of FFR or filter media. Previous studies evaluated meltblown polypropylene filter media using NaCl aerosol testing and found that filtration efficiency remained ≥95% after twenty treatment cycles consisting of 75°C for 30 min (14,17,20). On the other hand, Viscusi et al. (15) autoclaved two FFR models at 121°C/15 psi for 15 or 30 minutes and reported that filtration efficiency of both models decreased below minimum NIOSH requirements. A study investigating dry heat at lower temperature settings (110°C) reported that reduction of filtration efficiency was FFR model-specific (16). Previous studies also reported that VHP did not result in the reduction in the filtration efficiency of any of the N95 FFR models tested, while showing significant inactivation of spores, bacteriophages, and SARS-CoV-2 virus (3,9,14,16,20). Battelle CCDS (Critical Care Decontamination System)™ was approved by the FDA as the first emergency use authorization (EUA) for decontamination system for FFRs (3). In Battelle’s system, the 3M 1860 model (RM) was shown to maintain filtration performance for 50 treatment cycles of VHP treatment using the Clarus® R HPV generator (utilizing 30% H₂O₂) (3). Additionally, FFR fit was shown to be unaffected for up to twenty VHP treatments cycles using a manikin headform which was assessed on one FFR model without donning and doffing between cycles.Strap degradation, however, did occur after twenty treatment cycles (3). In addition, the FDA has issued an EUA for the SSS VHP N95 Respirator Decontamination System manufactured by Stryker Sustainability Solutions where reprocessing is limited to three cycles (2).

Fischer et al. (14) incorporated test subject quantitative respirator fit testing of intact FFRs using a PortaCount® following each cycle of heat treatment and wear for two hours and reported that the mean fit factor of six fit tests remained >100 for both one and two treatment cycles, but fell slightly below 100 after three dry heat (70°C) cycles. They also reported that one FFR did not have any physical deformation. The same study reported that the mean fit factor of six fit tests remained acceptable (according to OSHA criterion) following three VHP treatment cycles.

We also collected additional measures related to the physical appearance, physical comfort, and user acceptance of decontaminated FFRs/MMs. The purpose of these measures was to determine whether different decontamination methods impact FFRs and MMs in ways beyond physical fit, which could affect perceptions of healthcare personnel in terms of their use of decontaminated devices and ultimately if the use of decontaminated devices would affect HCPs’ performance and delivery of care. We plan to analyze and report upon this component of the study in a future publication.

There were limitations to this study. We only tested 3 types of FFRs, yet there are approximately 500 N95 FFR models approved by NIOSH and many medical masks around the world. Each FFR, MM, or CM model has distinct filter/strap/ear loop/tie materials and design characteristics (e.g., surface area, dead space, nose padding). The differences in materials and design characteristics among all available FFR/MM/CM models result in variable outcomes for decontamination efficacy and effect of decontamination on FFR performance. CDC recommends that a decontamination method’s effectiveness be evaluated for specific FFR models in collaboration with the manufacturer, and if needed, a third-party laboratory (21).

Another limitation within our study was that our primary inactivation studies were performed with direct application of a fixed quantity of MB onto coupons and only two of our experiments tested the spraying of MB onto a whole mask or the soaking of PPE material into a solution of MB. We attempted to normalize the volume of MB applied to the coupons to the amount of per square surface area. For example, if 7-8 mL of MB solution was sprayed onto an intact mask (4 sprays on outer and 2 sprays on inner surface), based on the surface area of a 3M N95 1860 respirator (RM), this would result in 10-30 ul of MB for the each 1 cm² coupon.

In our clinical specimen experiment, we demonstrated SARS-CoV-2 inactivation with MBL. Obtaining whole masks worn by healthcare personnel who cared for COVID-19 patients was not in the scope of our study, yet would add strength to our finding that a saliva specimen from a COVID-19 positive patient was inactivated by the MBL treatment.

All labs performed their studies with ‘biological triplicates’ – three like-masks or respirators from the same respective manufacturing batches were used for every mask tested. We attempted to enhance and strengthen our findings by leveraging four different virology testing sites using three different coronavirus species, including two surrogate coronaviruses. Our findings with the surrogate coronaviruses allow future studies to test diagnostics and therapeutics at lower levels of biocontainment than required for SARS-CoV-2. We believe that conducting our simultaneous experiments across two continents using the same methodology but in different labs, as well as conducting additional experiments using a SARS-CoV-2 clinical sample provides generalizable knowledge. The heterogeneity of light administration methods replicated practical light scenarios. The ovens used in the studies were all different and only standardized for the temperature administered (75°C). The use of alternative heat and alternative light producing methods parallels what we would expect in the real-world settings – no standard products, but ability to standardize a protocol.

The biocompatibility of MB on masks was not tested in this study yet the concentrations of MB – 10 μ M – used to spray onto the masks or used to drop onto mask coupons, were far below the clinically used MB concentrations administered intravenously for methemoglobinemia (1 mg/kg) (5), orally as third line therapy for malaria (22), or applied intranasally for Staphylococcus aureus (23). The amount of MB sprayed onto each mask using our 10 µM solution was <0.026 mg. Our team will be testing the amount of MB that a wearer might inhale during the course of an 8-10 hr shift wearing a mask. Although human fit testers of the MBL-treated masks did not experience any negative effects, more rigorous inhalational biocompatibility tests need to be done.

We did not replicate multiple real-life PPE donning and doffing over multiple shifts or extended wear which could affect FFR fit and performance over time. Degesys et al. tested the real time fit using models we used in our study. The 3M 1860 maintained fit for more than 5 donnings (24). In addition, off-gassing of decontamination chemicals was not evaluated although anecdotally, some wearers of the VHP+O₃ treated masks complained of an “acrid” or “acidic” smell while some wearers of the MBL-treated masked claimed that there was a “mild” and “calming” aroma appreciated. Human factors assessments, including user acceptance of decontaminated FFRs, MMs, and CMs by HCPs are underway. Further testing of the MBL method for user acceptance and workplace compatibility are also being explored.

We demonstrate that MB activated by broad-spectrum white or red light effectively inactivates coronaviruses including the novel SARS-CoV-2 on typical medical mask surfaces without affecting mask material integrity, or alterating mask fit and filtration characteristics. This series of studies provides evidence that MB can be applied to the surface of a FFR or MM and illuminated by a light source as a disinfection technique, enabling safer reuse of single-use disposable respirators and masks in shortage situations. One could envision using this decontamination process after each work shift in a healthcare setting. Integrity was maintained for the FFR models tested in this study after 5CD, however, the variability in the the design and materials (e.g., straps and earloops) may still lead to mask performance changes with successive donnings and doffings that will still need to be assessed to ensure adequate fit of the respirators. Residual MB remaining on the mask surface activated by ambient white light could also provide a novel means of continual disinfection of viral particles which become adherent to the mask surface in-use. Indeed, under ambient lighting condition, the surface of the masks treated with MB could produce up to 3.2×10¹⁵ molecules of singlet oxygen per second (Supplemental Table S6). Other potential uses include application of MB to face coverings used in low resource medical and general community settings, in order to enhance protection, not only of persons in proximity to the face covering user, but also to increase protection of the user from airborne virus. In the medical setting, on site, MBL spot treatments of PPE touch points prior to doffing may also decrease accidental contamination and may potentially provide increased protection of healthcare and laboratory workers. A further series of experiments intended to support the implementation of these additional uses of MB and light are underway.