Abstract
During the COVID-19 pandemic the use of cloth masks has increased dramatically due to the shortage of medical masks. However, the efficiency of this material is controversial. We aimed to investigate the efficiency of cloth masks in reducing transmission and contamination by droplets and aerosols for the general population and healthcare workers. Electronic databases were searched without year or language restrictions. Clinical and laboratorial studies were included. The risk of bias (RoB) was assessed using an adapted quality checklist for laboratory-based studies. ROBINS-I tool and Cochrane RoB 2.0 were used to evaluate non-randomized (n-RCT) and randomized clinical trials (RCT), respectively. The quality of the evidence was assessed through GRADE tool. From the eleven studies selected, eight were laboratory-based studies, one non-randomized and one RCT supported by laboratory data. Between the evaluated fabrics only three presented a filtration efficiency >90%. Hybrid of cotton/chiffon (95%CI 95.2 to 98.8), hybrid of cotton/silk (95%CI 92.2 to 95.8) and cotton quilt (95%CI 94.2 to 97.8). A meta-analysis was not feasible due to a high methodological heterogeneity. The overall quality of evidence ranged from very low to moderate. Despite the lower efficiency compared to medical masks, laboratorial results may underestimate the efficiency of cloth masks in real life. Cloth mask efficiency is higher when made of hybrid fabrics (cotton/chiffon, cotton/silk) and cotton quilt, mainly with multiple layers. In pandemic situations any measure that can contribute to source control at the population level can have a beneficial effect. However, cloth masks are not recommended for healthcare workers.
INTRODUCTION
According to the World Health Organization (WHO)1, viral diseases continue to emerge and represent a serious issue to public health. In the past few months, the COVID-19 pandemic has been the focus in scientific journals and the media. Frequent handwashing, barrier measures such as gloves, gowns and masks and isolation of suspected cases are some of the recommended procedures to reduce transmission in respiratory diseases2. Knowing COVID-19 is highly contagious, some experts and countries have encouraged or even implemented mandatory facial covering in public as a form of prevention3.
Recent studies4,5 reported that viral shedding of patients with the SARS-CoV-2 was higher at the time or before symptom onset. It means that a considerable portion of infected individuals with the new coronavirus are asymptomatic or pre-symptomatic patients and can transmit the virus during routine activities like speaking, coughing, or sneezing. Therefore, it is extremely important that everyone use masks, even those who did not present any symptoms.
Surgical masks, N95 respirators and similar are effective barriers that can help preventing COVID-19. However, due to the shortage of these products at the market6 it only should be used by healthcare workers. For the general population, the Centers for Disease Control and Prevention7 recommends wearing cloth mask covering in public settings, which are a simple and low coast measure that may have a big social impact.
The main objective of use these masks in public is to decrease transmission by pre-symptomatic infected individuals who continue to move freely. This is known as source control and refers to the effectiveness of blocking droplets from an infected person, when droplets expelled are not small enough to squeeze through the weave of a cotton mask8. When facing an epidemic episode, authorities must decide on the best actions to reduce the social impact, and source control becomes a critical matter in the debate about whether the public should wear masks9.
Studies suggested that in laboratory settings cloth masks are less effective than surgical masks10,11, but they seemed better than no protection at all12. Although some studies showed limited evidence regarding the use of face masks, absence of evidence is not evidence of absence13. With no effective vaccine or treatment, reducing the rate of infection is a urgency14.
Public policy makers need urgent guidance on the use of masks by the general population as a tool in combating SARS-CoV-2, based on the best available evidence. Therefore, the aim of this systematic review was to evaluate the existing evidence on the efficiency of homemade or commercial cloth masks compared to surgical masks and N95/others respirators in reducing transmission and contamination by droplets and aerosols in the general population and among healthcare workers.
MATERIAL AND METHOD
Search strategy and selection criteria
The present systematic review was registered in the PROSPERO database (https://www.crd.york.ac.uk/PROSPERO, CRD42020178417) and was reported according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines (www.prisma-statement.org)15.Electronic searches, without date and language restrictions, were carried out according to the PECO strategy. Eight databases were used: PubMed, Scopus, Web of science, Cochrane, VHS, OpenGrey, Google Scholar and Clinical Trials (Table 1). An additional manual search was carried out in the reference list of the included articles, as well as in some hand-picked electronic journals. Alerts were received by April 30, 2020.The eligibility criteria were defined according to the PECO research strategy for clarity in resolving the question: Can homemade or commercial cloth masks be used instead of surgical masks and N95 respirators as an alternative in reducing transmission and contamination by droplets and aerosols for general population and healthcare workers?
Inclusion Criteria:
Problem: Droplet and/or aerosol dispersion contamination.
Exposure: Homemade and/or commercial cloth masks.
Comparison: Surgical mask and/or N95 respirator.
Outcome: The efficiency of handmade or commercial cloth masks in reducing contamination and the transmission of contaminated droplets and aerosols, by means of laboratory and clinical tests that use surgical masks or N95 respirators for comparison.
Study Types: Randomized or non-randomized clinical trials, observational and laboratory studies.
Exclusion Criteria:
No comparison group.
Case series, opinion articles, animal studies and narrative reviews.
Data collection
Two authors, independently, sorted the articles by title, abstract and full text, using the bibliographic reference manager Endnote (version X7, Thomson Reuters). Disagreements during study selection and data extraction were settled through a consensus meeting and, when appropriate, by consulting with a third author. The qualitative data extraction table included the following information: Author, Year and Country; Exposure; Comparison; Sample and Method and Authors’ conclusions.
Data analysis
For all laboratory-based studies evaluated in this systematic review, the Checklist for Quasi-Experimental Studies (non-randomized experimental studies) from The Joanna Briggs Institute (JBI)16 was used. The checklist was adapted according to the statements proposed by CRIS Guidelines (Checklist for Reporting In-Vitro Studies)17, which suggests evaluating factors such as the randomization process, blinding and statistical analysis. The evaluated criteria were divided into seven domains which were categorized with “yes”, “no” or “unclear”. The checklist was individually analyzed for each study and classified as low, moderate or high risk of bias. This final classification was assigned according to the number of domains that presented “no” or “unclear” as an answer. One or two domains were considered as low risk; three or four as moderate risk; and five or more as high risk of bias.
For the evaluation of RoB for the non-randomized clinical trials, the ROBINS-I-tool18 was used. The evaluated criteria were divided into pre-intervention, intervention, and post-intervention categories. The RoB was classified as low (one or two domains with “moderate” or “high”), moderate (three or four domains), serious (five or six domains), critical (all the domains), and no information accordingly19.
For the randomized clinical trial, the RoB was performed using the Cochrane Collaboration RoB 2.0 tool20 which uses the following domains: random sequence generation, allocation concealment, blinding of patients and personnel, blinding of outcome assessor, incomplete outcome data, and selective outcome reporting. Low risk of bias was considered when all key domains were considered at low risk; unclear risk of bias was considered when one or more key domains were unclear and high risk of bias was considered when one or more key domains were considered at high risk.
A meta-analysis was not feasible due to the high methodological heterogeneity identified; however, a detailed qualitative synthesis was performed. and the quality of the evidence of the included studies was performed using GRADE (Grading of Recommendations Assessment, Development and Evaluation – https://gradepro.org/)21. The following outcomes were analyzed: Filtration efficiency (%), penetration level (%), airflow resistance, protection factor, cough experiment, pressure drop, surface masks test and occupational health which includes clinical respiratory illness (CRI), influenza-like illness (ILI), laboratory-confirmed respiratory virus infection and pressure differential.
RESULTS
A total of 2047 records were initially identified in the 8 electronic databases searched: PubMed (n=898), Scopus (n=9), Web of Science (n=7), Cochrane (n=279), Virtual Healthy Library (n=249), OpenGrey (n=2), Google scholar (n=600) and Clinical Trials (n=3). After the removal of 218 duplicates through the Endnote manager, 1829 titles and abstracts were examined. Fifteen records which satisfied the inclusion/exclusion criteria were retained for full text assessment. From those 15 studies, six were excluded: one did not define the type of masks compared22; one did not report how the particle penetration rate of the masks compared was obtained23; two records examined only factors influencing compliance with the use of medical and cloth masks amongst hospital workers24,25; and two evaluated the effectiveness of cloth masks after washing without any comparisons26,27 (Table 2).
One additional article was identified after hand search and another was found through a search alert. Finally, 11 articles were selected and included in the qualitative synthesis of this systematic review11,12,19,28–35 (Figure 1). The summaries of qualitative and quantitative data are shown in Table 3 and Table 4 respectively. Attempts to communicate by email with corresponding authors were made when data were unavailable. However, only one author responded19.
According to each type of study, a different tool for assessing RoB was used. From the 11 selected studies for qualitative analysis, there were nine laboratory studies, one non-randomized clinical trial and one randomized clinical trial complemented by laboratory data. RoB was performed separately for each outcome within each study.
An adapted JBI checklist for Quasi-Experimental Studies (experimental studies without random allocation) was applied to ten studies11,12,19,28–33,35. Seven domains were evaluated: randomization processes; clearly described methods, interventions, outcome measures; blinding of the assessments; reliable measurement of outcomes and proper statistical analysis (Table 5).
Only three studies19,29,30 reported on the randomization process and one19 informed the blinding process, but this investigation was classified as lacking clear information about reliable measurement of outcomes.
Five studies12,19,30,32,35 were classified with a low RoB, but only one30 reported correctly all the domains, excluding the blinding of the assessments that was unclear. Four studies11,29,31,33 were classified with a moderate RoB mainly for not reporting any randomization process and for not clearly describing other domains. Only one study presented a high RoB28 because it reported only on the reliable measurement of outcomes and on the interventions.
The ROBINS-I-Tool (Risk of Bias in Non-randomized Studies-of Interventions) was used in one study34 that was classified with a high risk of bias (Table 6). The major reason for this RoB rating was due to bias in selection of participants, who had been invited to participate in the research; and bias in classifying interventions since it did not report if the cough velocity was measured and if the patients were under treatment, which can be confounders since the cough velocity and the use of medications can modify the results. In addition, no inclusion and exclusion criteria of participants had been established and this can lead to a heterogeneous sample and unrealistic results. For the cluster randomized trial19, RoB was evaluated according to the Cochrane collaboration RoB 2.0 tool, and was rated as low in all domains: random sequence generation, allocation concealment, blinding of patients and personnel, blinding of outcome assessor, incomplete outcome data, and selective outcome reporting (Table 7).
Regarding the material used in the face masks, six studies evaluated several household materials that could possibly be used for making cloth masks12,28,29,32,33,35, while four articles evaluated only commercially available cloth masks11,19,30,31. Four studies compared cloth masks with surgical masks only12,28,31,34, one compared them with N9529 respirators only, one with N95 and FFP-2 respirators30, four studies compared cloth masks with both surgical masks and N95 respirators or similar (FFP-2)11,32,33,35, and only one study compared cloth masks with surgical masks, and FFP-2 and FFP-3 respirators19 (Table 3).
Regarding the experimental model, performed by simulation, five studies used NaCl aerosol with particles size reported: 0.075μm29,31,33, 1μm29, 0.02 to 2μm19 and 10nm to 10 μm35. The air flow speed in these studies was 33L/min and 99L/min29, 85L/min31,33, 95L/min19 and in the last study two velocities 35 L/min and 90 L/min41 were used. Two studies used microbial aerosols: one was contaminated aerosol with Staphylococcus aureus of 1μm in diameter at 28L/min air flow speed28, one used Bacillus atrophaeus with 0.95-1.25μm and Bacteriophage MS2 with 0.023μm, both with a 30L/min air flow speed12. One study11 used Polystyrene latex and Diesel Particles from 0.03 to 2.5μm in an air flow speed of 8 and 19L/min; and one study30 used high particulate matter from 0.1 to 2.5μm with 40 and 80L/min. Three studies evaluated more than one outcome, and also used volunteers12,19,32. One study34 did not perform an experiment by simulation and concluded that both surgical and cloth masks are inefficient in containing the spread.
Anti-contamination measurements
For anti-contamination measurements, the filtration efficiency (%) was evaluated by six studies11,12,19,28,31,35, where three of them compared cloth masks with surgical masks only12,28,31. The first one12 analyzed several homemade cloth masks and found that better results were achieved by the tea towel with 2 layers (96.71 ± 8.73) and vacuum cleaner bag (94.35 ± 35), with results similar to surgical masks (96.35 ± 0.68). Cotton mix (74.60 ± 11.17) and 100% cotton T-shirts with 2 layers (70.66 ± 6.83) also presented good results, while linen (60.00±11.18) and silk (58.00 ± 2.75) had the worst results. The second one12,28 analyzed three different cloth masks and reported that the best result was achieved by the twill cloth mask(93.6±1.16) with no difference compared to Hopes® surgical mask (98.1±1.02, p<0.05), and the last one31 concluded that the surgical mask with two filter screens presented 60-80% of filtration efficiency while cloth masks about 20%.
Three studies11,19,35 compared cloth masks with both surgical masks and N95 respirators or similar. The first one11 reported that the efficiency of cloth masks presented the worst results (39% to 65%) in comparison to the other two groups, the second study19 noted penetration of particles through the cloth masks to be very high (97%), but neither study reported the fabric of the cloth masks. The last one35 found that hybrid fabrics potentially provide protection against the transmission of aerosol particles, with a filtration efficiency of three types of hybrid fabrics: cotton/chiffon (97 ± 1), cotton/silk (94 ± 2) with no gap, and cotton/flannel (95 ± 2) even better than N95 respirators (85±15) in relation to <300 nm particles.
The penetration level (%) was measured by three studies29,30,33. The first one33 compared cloth masks with surgical masks and N95 respirators and found a high penetration level in handkerchiefs mainly made of gauze and with one layer (99.57 ± 0.40), and better results were found in a certified N95 respirator group with a penetration level (0.62 ± 0.36). The remaining two29,30 compared the cloth masks only with N95 and/or FFP-2 respirators and both of them noted a high penetration level in cloth masks in relation to comparison group. One29 found better results between the cloth fabrics in sweatshirts with 40% of penetration level at 33L/min and 57% at 99L/min. The other30 found that a mask named Yi Jie PM2.5 presented the lowest degree of penetration between the other cloth masks options with 67.3% (IQR: 56.6%,75.2%), but even so with worse results when compared to the best N95 brand 3M9322 1.8% (IQR: 0.6%,4.7%).
Recommended Cloth Masks based on 95% Confidence Interval >60% for Filtration Efficiency and <10% for Penetration Level (Particles Size <0.03 μm) are reported in Table 4.
Occupational health was evaluated by only one study19 and the rates of clinical respiratory illness (CRI), influenza-like illness (ILI) and laboratory-confirmed virus infections were higher in the cloth mask arm compared to medical masks, mainly ILI, with a relative risk =13.00 (95% CI 1.69,100.07).
Protection factor32 showed that surgical masks provided about twice as much protection as homemade masks, FFP2 masks provided 50 times as much protection as homemade masks, and 25 times as much protection as surgical masks.
Anti-transmission measurements
The protection factor of cloth, surgical and FFP-2 masks were evaluated by one study32 which showed that cloth masks presented a considerably lower protection factor (1.9, CI95% 1.5,2.3) especially in children. Protection offered by a surgical mask and FFP2 respirator did not differ.
Two studies12,34 evaluated particle dissemination when coughing. The first one12 found that both surgical and cloth masks reduced the total number of microorganisms expelled when coughing in comparison with coughing without a mask, while the second one34 found that neither cloth or surgical masks effectively filtered the virus expelled when coughing.
Breathability
Studies evaluating pressure drop (PD)12,33 and airflow resistance (Pa)28,31 had demonstrated that tea towel12, vacuum cleaner bag masks12, cotton handkerchief with four layers33, twill weave28 and bleached cotton28 had greater potential to block contaminated patient particles outside the cloth mask. However, they can cause a suffocating sensation to the user.
On the other hand, some of the evaluated fabrics presented a good breathability, such as calico28, silk12, linen12, cotton and gauze handkerchief33. One study35 reported that the average differential pressure across all of the fabrics studied at a flow rate of 1.2 CFM was found to be 2.5 (0.4) Pa, indicating conditions for good breathability, but we can’t claim that these cloth masks are able to contain or reduce particles expelled by the user.
Quality of the evidence
GRADE assessment was divided into anti-contamination and anti-transmission and breathability outcomes. For the outcomes included in the anti-contamination the quality of the evidence ranged from low to moderate level (Occupational health). For the anti-transmission and breathability outcomes, the quality of the evidence ranged from very low to moderate due to the bias of the included studies and magnitude of effect (Tables 8 and 9).
In general terms, surgical and N95 and/or FFP-2 masks presented better results in most of the factors evaluated in comparison with the cloth masks, with very low to moderate certainty of evidence level depending on the outcome analyzed. Regardless of some benefits for cloth mask users, the results are hard to summarize and generalize because of the variety of fabrics and layers evaluated.
These results should be viewed with caution given the quality of the evidence and the fact that almost all the included studies evaluated the outcome of interest in a laboratory setting. Furthermore, elements of statistical precision between the groups are scarce, and outcomes such as degree of protection, pressure drop, surface masks test and occupational health were each evaluated in only one study.
DISCUSSION
Our results suggest that cloth masks present worse outcomes for filtration efficiency, penetration level and protection factor in comparison with medical masks, when evaluated in a laboratory-based examining small particles. In accordance with other study8, these results seem to substantially underestimate the efficiency of cloth masks for source control in real life when referring to blocking droplets ejected by the wearer, since in most cases the particles used in these studies were smaller than a droplet and generally ranged between 5 μm to 10 μm36,37. For this reason, it is suggested that the use of cloth masks by the general public is likely a useful public health measure in reducing COVID-19 contamination and transmission. In addition, the fact that cloth masks are not as effective as surgical masks does not mean that they provide no protection. Anything that contributes to controlling the spread of a virus should be encouraged from a population-based point of view. Multiple approaches that alone do not a have a major impact when combined could have a multiplicative effect in slowing the spread of a virus like COVID-19 by reducing the transmission rate.
Overall, the filtration efficiency of the fabric depends on a variety of factors: the composition of the fabric and some characteristics of the particles to which it is exposed such as their size and velocity. These factors are fundamental to evaluate the quality of the masks. Only seven studies11,12,29–33 presented particle sizes compatible with the new coronaviruses (0.06-0.14 μm)38. This lack of complete information directly affected the potential bias of these studies. If the particle sizes were known, we could have better evaluated the efficiency of cloth masks for the general population against the coronavirus.
Only one study34 evaluated the use of cloth masks in patients with COVID-19, but it is important to clarify that this study presented some limitations such as a small sample size and inconsistent data, including no detection of viral load in one participant’s cough test (including without a mask) and no detection of viral load in the inner surface of the masks in three of four patients after coughing. The other studies assessed other types of bacteria and viruses, but this did not seem to affect the results. The study that evaluated the use of cloth masks by healthcare workers19 did not recommend their use by these professionals. A recent systematic review39 showed that low quality evidence was presented in studies evaluating the use of PPE, face masks (surgical and N95) and eye protection to prevent infectious diseases in healthcare workers. The authors highlighted the urgent need for randomized clinical trials with better methodological quality. However, results in a healthcare setting are not readily generalizable to the population where any measure, even not as efficient as a measure in a healthcare setting, can provide some source control.
Another recent systematic review14 investigated physical distancing, face masks and eye protection to prevent person-to-person transmission of COVID-19 and they supported that physical distancing of at least 1 m is strongly associated with protection, but distances up to 2 mm might be more effective. Regarding the use of face masks, it was found that it could result in a large reduction in risk of infection, with stronger association with N95 or similar respirators when compared with surgical masks. Eye protection also was associated with less infection.
Among the studies that reported the materials used to fabricate cloth masks, the vacuum cleaner bag presented good results, but it’s important to clarify that this material has a high pressure drop, rendering it unsuitable for a face mask, therefore the use of tea towel was recommended instead12. In relation to layers of masks, the number used seems to be directly proportional to the filtration capacity in most of the laboratory studies and could be a solution to improve the results achieved by cloth masks40–43. The combination of various commonly available fabrics can potentially provide significant protection against the contamination of aerosol particles, as a hybrid of cotton and silk mask seems to present results of filtration efficiency very similar to surgical and N95 masks35.
Cloth masks can be effective depending on the fabric and number of layers used. It could decrease the air passage from inside to outside of the masks, thereby favoring the decrease of the microorganisms expelled during speaking, coughing, or sneezing. However, it is critical that it is well adapted to the facial contour, since the presence of gaps caused by an improper fit of the mask can result in over a 60% decrease of their filtration efficiency35. In addition, some authors recommended that in situations of public emergency, with limited evidence, mechanistic and analogous evidence and professional judgment become important. In these cases, the use of facial masks, along with other health measures, such as personal hygiene, can help mitigate the COVID-19 epidemic44.
A recent rapid systematic review45 evaluated the use of medically manufactured facemasks and similar barriers to prevent respiratory illness such as COVID-19. According to the RCTs, the results showed that the use of facial masks may present little protection against primary infections through casual contact with the community, and modestly protect against domestic infections when infected and non-infected members wear face masks. In observational studies the evidence in favors of wearing facemasks was stronger. This is an important point to be cited since the clinical studies could often suffer from poor compliance and controls using facemasks45. Therefore, the correct and continuous use of these protections by the public could improve the clinical results.
This fact can also be supported by a mathematical modelling study that described the spread of COVID-10 infection. Modelling studies suggest that if most people wear masks, the transmission rate can decrease to 1.09. Moreover, cloth masks could be an additional tool to enhance awareness of the importance of physical distancing in public places, serving as a visual reminder8.
Another study46 discussed the potential effectiveness of the universal adoption of homemade cloth facemasks. They found that that the growth rate of deaths in countries without mask norms is 21%, while in countries with such norms is 11%. Although researchers may disagree on the magnitude of the reduction in transmissibility, the benefits found can be highly expressive and beneficial to the transmission and control process of the disease8. Public use of facemask may increase awareness regarding the disease among the population and can contribute to the reduction of the transmission rates.
It is well-known that the virus may survive on the surface of face masks47, and contamination may occur since the cloth mask may transfer pathogen to bare hands during the repeated donning and doffing17, so it’s very important to wash hands as much as possible and wash the masks daily. Conversely, a study21 showed that washing and drying practices could drop by 20% the filtering efficiency of cloth masks after the 4th cycle, due to the increase of the pore size and the expansion of the fabric. It is important to highlight that the masks were air dried to make sure that the cloth fibers were not stretched out, since stretching cloth masks surface also altered the pore size and potentially decreased the filtering efficiency. Further studies about wash and dry care of cloth masks are needed to obtain a longer durability with efficiency. Moreover, authorities need to provide clear guidelines for the use, cleaning, and reuse of facemasks.
A guideline from the WHO48 encourage the use of respiratory hygiene in all people with acute respiratory infections (ARIs) and it includes the use of medical or cloth masks. Although the quality of evidence has been considered very low2, there was consensus that the advantages of the use of respiratory hygiene and an assessment of values and preferences provided sufficient basis for the strong recommendation. Thus, the importance of cloth masks use by the general population seems an effective way of source control, as people in a pre-symptomatic phase can already spread the virus. This is a simple and low-cost measure that in conjunction with other strategies can be extremely helpful in control and mitigation of the disease.
This systematic review identified some limitations in the primary studies and only two of the included studies were clinical trials. A more realistic comparison between groups was hampered by the lack of detailed features information on the masks studied. There was a lack of studies comparatively assessing the various fabrics, utilizing different particles sizes and designs of cloth masks, and taking into consideration the importance of a good fit on effectiveness. Besides that, new studies with a better methodological quality and randomized clinical trial specifically related to COVID-19 are in urgent needed.
CONCLUSION
Cloth masks seem to provide some degree of protection against contamination and transmission by droplets and aerosols. It is suggested that the use of cloth masks by the public is a useful public health measure that can protect the wearer and at the same time act as source decrease disease transmission. However, according to one RCT, cloth masks should not be recommended for healthcare workers. Based on very low to moderate quality of evidence the level of efficiency of cloth masks is difficult to generalize because of the variety of fabrics and layers evaluated, the efficiency is higher when cloth masks are made of hybrid fabrics with multiple layers.
Data Availability
Not applicable.
Acknowledgments
We thank Dr. Sandra Meilhubers for comments and pre-submission review. There was no funding source for this study.