Effect of gamma irradiation on filtering facepiece respirators and SARS-CoV-2 detection

To cope with the shortage of filtering facepiece respirators (FFRs) during the coronavirus (COVID-19) pandemic, healthcare institutions were forced to reuse FFRs after applying different decontamination methods including gamma-irradiation (GIR). The aim of this study was to evaluate the effect of GIR on the filtration efficiency (FE) of FFRs and on SARS-CoV-2 detection. The FE of 2 FFRs types (KN95 and N95-3 M masks) was assessed at different particle sizes (0.3–5 µm) following GIR (0–15 kGy) delivered at either typical (1.65 kGy/h) or low (0.5088 kGy/h) dose rates. The detection of two SARS-CoV-2 RNA genes (E and RdRp4) following GIR (0–50 kGy) was carried out using RT-qPCR assay. Both masks showed an overall significant (P < 0.001) reduction in FE with increased GIR doses. No significant differences were observed between GIR dose rates on FE. The GIR exhibited significant increases (P ≤ 0.001) in the cycle threshold values (ΔCt) of both genes, with no detection following high doses. In conclusion, complete degradation of SARS-CoV-2 RNA can be achieved by high GIR (≥ 30 kGy), suggesting its potential use in FFRs decontamination. However, GIR exhibited adverse effects on FE in dose- and particle size-dependent manners, rendering its use to decontaminate FFRs debatable.

The filtering facepiece respirators (FFRs), including N95 masks, play an important role in infection prevention and control by reducing the airborne transmission of infectious illnesses from patients to health practitioners 1,2 . The N95 mask is a high efficacy single-use FFR that can block up to 95% of particles of ≥ 0.3 µm in diameter. The N95 mask is approved by the United States Food and Drug Administration (FDA) and the National Institute for Occupational Safety and Health (NIOSH) for use as personal protective equipment (PPEs) in medical settings 1,2 . Due to the global shortage of N95 masks during the outbreak of Coronavirus Disease (COVID-19) caused by the emerging severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), NIOSH has issued a recommended guidance for the extended use and limited reuse of N95 masks in emergency medical settings 3 . Although the NIOSH guidance can significantly reduce the consumption of N95 masks during the pandemic, concerns about these policies have been raised 4,5 . These concerns are related to the long survival of the SARS-CoV-2 virus on the outer layer of N95 masks, leading to cross-contamination between staff and patients 4,5 . Therefore, decontamination and subsequent reuse of the N95 masks were highly recommended to cope with the expected shortage in such a pandemic.
There are six well-characterized N95 decontamination procedures including vapor hydrogen peroxide (VHP), ethylene oxide, moist heat incubation, microwave oven, ultraviolet germicidal irradiation (UVGI), and gamma irradiation (GIR) [6][7][8][9][10] . Although the USA Centers for Disease Control (CDC) did not approve the routine decontamination of N95 masks, it released emergency guidelines on the N95 decontamination methods and indicated that UVGI, VHP and moist heat had shown the most promising results 7 . However, these three methods have certain logistical and technical limitations including the small capacity, the limited penetration and the high risk of pathogen cross-contamination should the used FFRs be removed from their biosafety container/bag and handled individually during the decontamination process [6][7][8][9][10] .
Hypothetically, GIR has several advantages among other decontamination methods such as better penetration, better certainty of sterility and independence from temperature and pressure conditions 11 . However, concerns have been raised regarding the adverse effect of GIR on the physical properties of FFRs materials, in particular
Taking both dose rates together, the KN95 masks following GIR showed higher FE (81.5-100%) than 3M-N95 masks (FE = 60.8-98.6%) (Fig. 2). The FE of both masks at large PSs (2 and 5 µm) was somewhat stable compared to FE at small PSs (0.3-1 µm). In general, the FE values were decreased with the increased GIR doses. The highest decreases in FE of KN95 and N95-3M masks were observed following 15 kGy, with percentage decreases ranging between 0.60 and 22.8% (Supplementary Table S1).
For both dose rates, the average FE values of KN95 masks irradiated with 4, 5, 7.5 and 15 kGy were significantly (P ≤ 0.03) lower than control (0 kGy), whereas the average FE of N95-3M masks following all GIR doses (0.5-15 kGy) were significantly lower than control (P ≤ 0.025). Interestingly, no significant differences (P > 0.05) were observed between typical and low dose rates GIR regarding their effects on FE of both masks.
In terms of evaluating the effect of low dose rate GIR on the physical structure of FFRs, no signs of visible changes or damages were observed. However, microscopic images (using ZEISS Axio Vert.A1 microscope, ZEISS, Germany) of the fiber layers of both masks were snapped following 0, 0.5, 5 and 15 k Gy. The images showed some unusual micro holes in the irradiated fabrics compared to the control (Fig. 3). These holes may have resulted from alterations in polymer homogeneity caused by GIR, such as polymer clustering, cracking, and degradation, which may explain the instability of both masks at small PSs ≤ 1 µm following GIR. SARS-CoV-2 RNA detection using RT-qPCR assay. SARS-CoV-2 RNA stability following incubation at room temperature (RT) versus − 80 °C. Results showed that the viral E Gene RNA was detected in all samples at mean cycle threshold (ΔCt) values between 19.3 and 22.9 (Fig. 4). Interestingly, incubation of SARS-CoV-2 RNA samples at RT, as compared to − 80 °C, for 48 and 96 h had no significant effect on RNA detectability (Ttest, two-tailed P-values = 0.713 and 0.467, respectively). In addition, there were no significant differences in Ct values between 0, 48 and 96 h (RT along with − 80 °C) of incubation time (P = 0.293, parametric one-way repeated measures analysis of variance "RM-ANOVA").
SARS-CoV-2 RNA stability following gamma irradiation. The effect of the GIR (0.5088 kGy/h) doses (0 to 50 kGy) on the detectability of SARS-CoV-2 RNA was assessed in 2 viral genes, envelope (E) and RNA dependent RNA polymerase 4 (RdRp4). Results of the respective ΔCt values are presented in Fig. 5. Representative images of RT-qPCR results for both genes are presented in Fig. 6. Obviously, the ΔCt detection threshold increased with increasing irradiation doses, indicating progressive radiation-induced degradation of viral RNA. For both   (Fig. 5-B).

Discussion
We have evaluated the FE of N95-3M and KN95 following GIR as a potential decontamination method for subsequent reuse of FFRs as a crisis strategy in case of N95 masks shortage 3 . The results showed that the KN95 mask retained an initial FE of ≥ 99.7% at a PS of 0.3 µm or larger, whereas the N95-3M mask showed a range of FE between 77.1% at 0.3 µm and 98.7% at 5 µm (Fig. 1). The overall means FE of KN95 masks were significantly higher (P = 0.022) than N95-3M masks. This agrees with a recent study showing an average KN95 FE of 96.7% at PS between 0.03 and 0.5 µm 33 . In reasonable agreement with our results, Cramer et al. found that control    6 . It is worth pointing out that the latter study used the same FE evaluation parameters assessment used in our study in terms of using the PM air as a source of the particles, using the same AeroTrak particle counter (TSI, Model 9306) as a particle counter and using the same air filter face velocity (0.4 m/s). The current study hypothesized that the adverse effect of GIR on FFRs could be minimized by using low dose rate GIR. This was suggested from the literature where a potential association between increasing GIR dose rate and decreasing FE 6,[16][17][18][19] . Furthermore, a recent consultancy meeting on "radiation effects on polymer materials" organized by The International Atomic Energy Agency (IAEA) recommended expanding the research on the GIR dose rate and its potential impact on minimizing polymer alterations, stating that "it was determined that published studies on dose rate effects on polymers are largely lacking" 34 . Our results showed that the FE of KN95 masks was almost stable at PSs of 2 and 5 µm following GIR of either typical or low (FE = 98.3-100.0%) dose rates (Fig. 2-A). However, a decrease in FE of KN95 masks at PSs between 0.3 and 1 µm was observed following typical or low dose rates, reaching 88.0% and 81.5%, respectively. For N95-3M masks, our results showed that the FE value at PSs of 5 µm (94.3-98.7%) was more stable than other PSs following typical or low dose rates GIR  www.nature.com/scientificreports/ ( Fig. 2-B). Overall, no significant differences (P > 0.05) were observed between typical and low dose rates GIR in terms of their effects on FE of each mask. This is the first study to investigate the effect of GIR on KN95 masks as well as the first to compare the effect of different dose rates of GIR on FE of FFRs. Only five studies, with two of them are preprinted manuscripts, investigated the effect of GIR on FFRs using different evaluation parameters including GIR dose rate, particle size, air filter face velocity, flow rate, particle counter and source of particles (Table 2) 6,[16][17][18][19] . In disagreement with our findings, Cramer and colleagues observed a 67.0%, 54.5% and 22.1% decrease in FE of N95-3M masks, at PSs of 0.3, 0.5 and 1 µm following 1 kGy (2.2 kGy/h) 6 . In comparison, there were only 9.9%, 10.1% and 6.9% for the typical dose rate and 7.7%, 11.0%, and 8.6% for the low dose rate, respectively (Supplementary Table S2) in our study. It is worthwhile noting that the GIR dose rate used was higher (2.2 kGy/h) than those used in our study. Another study also showed a ~ 50% decrease in FE at PS of 0.075 µm of two other FFRS types following a single dose of 25 kGy (1.8 kGy/h), a dose that was not examined in our study 19 . Furthermore, the study showed that GIR exhibited significant changes in electrostatic charges of the filtration layer within the FFRs.
Furthermore, three reports showed mild adverse GIR effect (< 50% decrease in FE) on FFRs, which are comparable with our findings [16][17][18] . The first preprint showed that the average FE of N95 masks at ≥ 0.3 µm was decreased by 29.6% following 15 kGy, while it was 12.9% in our results (Supplementary Table S1) 16 . The second preprint showed a ~ 28% decrease in FE at 0.3 µm following 15 kGy, while it was 19.75% in our findings 17 . However, the latter preprint showed no effect of GIR on FE at 5 µm, which differs from our findings as ~ 4.5% decreases in FE was observed (Supplementary Table S1). The third preprint showed that GIR exhibited a 3.7% decrease in FE at 1 µm following 15 kGy, while it was 11% (9.6-12.4) in our results 18 . Taken all the five studies together, GIR had an adverse effect on FE of FFRs although they showed variations in FE data. These variations may be explained by the influence of various factors on the FE assessment such as FE evaluation parameters, GIR dose rate and FFR types.
To try to minimize the adverse effect of GIR on the fabrics of FFEs, two suggestions related to irradiation conditions could help preserve FFRs efficiency. The first suggestion is to irradiate the FFRs in a free-oxygen (e.g. nitrogen or vacuum) container 34,35 . The structural destruction of the PP (the main component of the FFRs filter fabrics) following GIR is particularly noticeable in the presence of air due to the oxidative damage of the PP's structure 12 . However, applying this idea should take into consideration that the structural components of viruses, such as nucleic acids, protein and lipid, could be more resistant to radiation in the anoxic than aerobic conditions 36,37 . The second suggestion is to irradiate the FFEs at a low-temperature atmosphere to decrease PP cross-linking within FFEs during the irradiation process 38 . Although cold-irradiation processing may protect against PP damage, it can also limit damage to the pathogens, the target of the irradiation process 39,40 .
Although our observation did not notice any visible changes in the FFRs structure following GIR, microscopic changes were observed within masks' materials, which may explain the FE instability following GIR (Fig. 3). These changes were expected as several studies reported that GIR exhibited significant physical changes in the PP, the main component in the filtration layers of FFRs [12][13][14][15] . Therefore, further monitoring of the potential changes in physical structure and electrostatic charge status of decontaminated FFRs should be considered. In terms of evaluating the fit factor of the irradiated FFRs, previous studies showed that GIR has no effect on the fit factor of FFRs although it degrades FFRs' FE 6,16,17 . In addition, a recent IAEA report agreed with the latter studies and showed that 24 kGy has no effect on the fit factor on irradiated FFRs 41 .
Interestingly, results showed that the viral E Gene RNA was detected at RT up to 96 h, with no significant difference compared to − 80 °C incubation (Fig. 4). Our data also showed that E and RdRp4 SARS-CoV-2 genes (Fig. 5) could be constantly detected up to 25 kGy, and randomly at 30-40 kGy, but not at 50 kGy. Two reports so far evaluated the ability of GIR to inactivate SARS-CoV-2 using the median tissue culture infectious dose (TCID 50 ) assay 30,31 . Both studies found that the cultured active SARS-CoV-2 was completely inactivated by GIR at the dose of 10 kGy. However, Leung et al. found that the E and the nucleocapsid (NP) SARS-CoV-2 genes can still be detected following 50 kGy using RT-qPCR assay 30 . The large difference between the GIR dose needed to observe complete viral inactivation (10 kGy) and to achieve complete viral RNA degradation (50 kGy) may emanate from the inherent characteristics of each assay. Although SARS-CoV-2 genes were still detectable at 40 www.nature.com/scientificreports/ kGy by RT-qPCR, it does not necessarily mean that RNA was intact, and the viral particle of SARS-CoV-2 was still infectious. The influence of GIR on other biological factors rather than RNA, such as protein and lipids, seems limited as a study showed that the integrity of viral morphology and protein structures of coronaviruses were preserved following 10 kGy 42 .
In conclusion, the KN95 mask showed a higher initial FE than the 3M-N95 masks. Irradiated KN95 and N95-3M masks showed microstructural changes within masks' fabrics associated with dose-dependent substantial reductions (≤ 18.5%) in FE at small particle sizes (0.3-2 µm) and moderate reductions (≤ 4.5%) at large (5 µm) particle size, the size of the most suspected droplets implicated in COVID-19 transmission. GIR dose rate does not seem to be an influencing factor on the FE of irradiated FFRs. Incubation for 4-days at room temperature has no effect on SARS-CoV-2 detectability using RT-qPCR assay. The mean cycle threshold (ΔCt) of viral RNA detection increased with increasing GIR doses with an absence of detection at very high doses.

Materials and methods
Authors confirm that all methods were carried out in accordance with relevant guidelines and regulations.  [16][17][18][19] . The first (typical dose rate) GIR source was used to irradiate FFRs (0.5-15 kGy) only, while the second (low dose rate) GIR source was used to irradiate FFRs (0.5-15 kGy) and SARS-CoV-2 RNA samples (1-50 kGy). The GIR doses ranged from 0 to 50 kGy and the required time needed to deliver the doses extended up to 98.16 h for the highest dose.

Filtration efficiency (FE) assessment of filtering facepiece respirators (FFRs). Filtration effi-
ciency (FE) measurement. The FE measurement used in this study is an in-house method that was previously described 45 . Briefly, a custom-designed air duct was manufactured to measure the FE of different FFRs types using the PM air as a source of the measured particles. The overall dimensions of the air duct are 19-cm-long, 14-cm-wide and 12-cm-high. The air duct consists of two parts, head and tail, that can be tightly joined together by three mold bolts to squeeze a filter in a sandwich manner, with no air leak present between both parts. The head of the air duct is connected to an AeroTrak particle counter (TSI, Model 9306) that counts particles with sizes of 0.3, 0.5, 0.7, 1, 2 and 5 µm at a flow rate of 2.8 liters/minute (L/min). The tail of the air duct is connected to an electrical fan that flows the PM air through the air duct tail to give a face velocity of 0.4 m/s; measured by Velocicalc Air Velocity Meter 9545 (TSI, product ID# 9545-A).
Initial filtration efficiency (FE) of FFRs. A total of 10 FFRs of each type (N95-3M and KN95) were subjected to FE evaluation. The 10 FFRs will be used later as controls (0 kGy) for FFRs irradiated with either low dose GIR (5 FFRs) or typical dose rate GIR (5 FFRs).
The particle number concentration of the PM air was assessed at least 5 times before FFRs assessment. For each PS, FE was calculated using the following formula: Filtration efficiency (FE) of FFRs following gamma irradiation. For each source of GIR (typical and low dose rates), 5 FFRs of each type (KN95 and N95-3M) were exposed to 8 different radiation doses (0.5, 1, 2, 3, 4, 5, 7.5 and 15 kGy) with 0 kGy as a control. In total, 45 masks of each type of FFRs were used in the study.
Detection of SARS-CoV-2 RNA following gamma irradiation. The stability of 5 SARS-CoV-2 RNA samples against GIR doses (0-50 kGy) was evaluated using the RT-qPCR assay targeting the E and RdRp4 SARS-CoV-2 genes.