There’s recently been an active debate about face masks and their role in protecting people from infection. The CDC and WHO have both issued advisories suggesting that masks are unnecessary, except for those in high-risk healthcare scenarios. The US Surgeon General’s twitter account sent out a perplexingly casual warning that “seriously people, stop buying masks!” followed by the statement that they aren’t effective for preventing community transmission of the virus.

Within days of the announcement, the official position of the CDC, WHO, and US Surgeon General’s office shifted. New recommendations were issued: masks, when used properly, can protect against community transmission of the virus. What happened? Why did they suggest not using masks in the first place, and what drove the change?

A multitude of needs

One reason for the recommendation was to decrease panic. There’s a shortage of PPE, masks among them, and a constantly rising case burden. People’s tendency in this situation is to hoard resources and sell them on the black market - which makes finding PPE for frontline workers that much more difficult. It seems reasonable that the agencies felt that a direct suggestion to the public would be the most effective way to decrease public demand and to open up supply chains for hospitals.

On the face of it, noble motivations. Production shortages are still plaguing places like 3M. The lack of masks is so pervasive, and hospitals so concerned about a spike in demand, that healthcare workers with their own PPE stockpiles are being prevented from using them. The logic is that if one doctor or nurse starts wearing a mask, soon everyone wants to wear a mask, but there are no masks, so panic will ensue.

In places like New York City, where PPE is has been broadly deployed, healthcare workers are being advised to stretch what limited resources there are. Guidelines for how to do this, euphemistically referred to as “optimizing the supply,” have been released by the CDC. Reuse of respirators, where clinicians remove a respirator and save it for later use, is not recommended. On the other hand, extended use, where clinicians simply work their shifts with as few breaks as possible in order to prevent the use of more than one respirator, has been given the green light. 

Another possible reason for the shifting guidelines is that an improperly used mask can actually become more dangerous than not having one at all. Public health agencies often complain about the difficulty of changing risky behavior in vulnerable populations, and a campaign on proper mask use and disposal may have been more than the stretched-thin resources of the WHO could tolerate. The simpler solution was to recommend that people go without.

There was immense blowback against the WHO advocating the public forego masks, which led to the release of new guidelines that recommend mask wearing, but remind visitors to their website that if you do decide to wear a mask, it’s imperative that you know how to use it and how to get rid of it. 

“Wearing medical masks when not indicated may result in unnecessary costs and procurement burdens and create a false sense of security that can lead to the neglect of other essential measures, such as hand hygiene practices. Further, using a mask incorrectly may hamper its effectiveness”.

Reminds me of how my mom always told me - you’ve got to learn how to any task correctly - even when it’s something as seemingly simple as washing the floor. 

The reason for concern on the WHO’s part is that surgical masks act like a filter. They become coated with virions after use. Once they’ve been used, masks become a biohazard that must be disposed of with care. In healthcare and laboratory settings, masks are discarded into a biohazard waste stream that gets sterilized before leaving the facility. Since most people don’t have this ability in their homes and neighborhoods, precautions are necessary to ensure used masks don’t become vectors of the very disease from which they’re protecting the public. 

How effective are masks?

Much of the confusion about mask advisories comes down to the fact that scientists don’t have a fantastic grasp on how, exactly, the coronavirus spreads. Person-to-person transmission can happen in a couple of ways. One way is direct contact, where there’s physical exchange of fluids. This can be active, like through sharing food and drink, kissing, etc - or can be passive - like someone sneezing, and someone else being in the way of the spray. Then there’s indirect contact, which is through airborne transmission or surface contamination.  

Airborne transmission is what the mask advisories hinge on, and it’s the piece of the puzzle that scientists are most unsure of. It’s consistently agreed upon that droplets larger than 5um contain the virus - but that these droplets don’t travel very far. Because they’re heavy, they tend to follow newtonian dynamics - meaning they fall to the floor shortly after being expelled. Smaller droplets, below 5um, tend to behave differently. Smaller droplets are light enough that they can be caught on air currents, and can circulate for a long time before falling to the ground.

Data from the field vary widely. In one case study of influenza transmission, 50% of viruses isolated from a emergency room are airborne, which means they’re below the 4um size threshold that keeps them circulating for an extended period of time. One study looked at environmental contamination of various sites in Wuhan - including two hospitals and a public area where people congregated and found low concentrations of airborne SARS-CoV-2 genetic material in all test sites except for the crowded public space. 

In the case of MERS, caused by a distantly related coronavirus, then asymptomatic patients wouldn’t be able to transmit the virus. This study, done on a single individual, epitomizes the piecemeal state of our understanding. The data we have is limited because there aren’t a ton of resources to study asymptomatic carriers during outbreaks - there’s barely enough resources to take care of the people that are sick and dying. 

In the current SARS-CoV-2 epidemic, a team in Hong Kong, investigating hospital transmission from infected patients, failed to detect any aerosolized virus, but a group at the University of Nebraska team found viral RNA present at all hospital sites tested, but at low levels -  the highest concentrations found near personnel interacting closely with infected patients. However, they found very low levels of viral RNA and weren’t able to culture viruses from any of the sample. A few things remain unclear - how many viruses are necessary to initiate infection, and how to differentiate between the presence of RNA and the presence of active virus.

That has been another confounding issue for understanding the likelihood of transmission - it’s possible to isolate viral RNA from the environment, but few samples contain significant amounts of culturable viruses. The amount of virions necessary to activate a human infection is still up in the air - but one study that examined the stability of viruses on surfaces, used 10^5 viruses. They chose this dose because that’s the amount of virions necessary to infect 50% of tissue culture cells. Contrasted with the double digit copy numbers of viral genetic material recovered from patients and their hospital surroundings, it’s difficult to produce a model in which aerosolization of the viruses is a significant contributor to their spread.

These data suggest a few different things. The first is that our ability to accurately evaluate the viral load off of air sampling isn’t great. More sampling, of more patients, in more locations, will help us understand if this is the case.

It could also be the case that some individuals are more infectious than others - the so-called “superspreader,” observed in both SARS and MERS epidemics. Superspreader is a little bit of an overstatement, since the baseline is a single individual that generates five novel cases - but is one possible explanation for the flaccid results in the studies above. The current explanation is that superspreaders simply had more contacts, more ER visits, more doctor exposure, than those who didn’t. But given the variability in other studies, it’s clear that further work needs to be done in order to evaluate the degree to which a COVID-19 patient is able to contaminate the air around them. 

Extrapolating from Influenza

The flu is similar to COVID-19 in that it is a respiratory tract infection caused by a virion that binds a specific protein in order to enter a specific cell. Once inside that cell both types of viruses replicate, cause damage to their host cell, and spread throughout the body. The specific cell types targeted by the virus are different, and so it’s not possible to say with certainty that what has been observed for influenza is true for COVID-19, but in the absence of other data it’s at least a good starting point. 

A relationship between location of infection and ability to transmit the virus in both influenza and COVID-19 is bolstered by some things we know about the viruses. Recent work in ferrets showed that, although influenza A can infect ciliated cells in both the upper and lower respiratory tracts, that infections are transmitted predominantly through upper respiratory tract exhalations.

The study that shows this was conducted by using two strains of Influenza A that could be differentiated through a synonymous mutation. Ferrets (donors) were infected at two sites with the different viruses, and then exposed to other ferrets (recipients) for a few days. Recipient ferrets, upon showing symptoms of viral infection, were swabbed. The swabs were sequenced and examined for similarity to the initial strains that had been inoculated into the donor ferrets.

For the most part, they found that the predominant viruses in the recipient ferrets were the ones from the upper respiratory tract of the donor ferrets. Different strains (H/N mixtures) of Influenza A had different kinetics of infection, but in all cases, viral shedding was over and done with by six days post infection. Analysis of respiratory droplets from the ferrets themselves showed that viral RNA was predominantly present in the larger particles, greater than 4uM across. 

This size of droplet indicated a direct form of infection, where droplets from donor ferrets containing virus were inhaled by the recipient ferrets. Inhalation as the mode of transmission was suspected because throat swabs of recipient ferrets were positive for viral RNA before upper or lower respiratory tract cells were. This makes sense, as inhaled air moves quickly through the upper respiratory tract. The air shifts direction at the back of the nasal passage, at which point any large droplets that haven’t been filtered by hairs and mucous would continue moving in the same direction, landing on the back of the throat.  

MERS, on the other hand, is a coronavirus that jumped into humans from a camel zoonotic reservoir. In camels it predominantly infects the upper respiratory tract, as that’s the location where the necessary receptor and surface molecules are expressed at the highest level. In humans, the same receptor/surface molecule combination is most expressed in the deepest parts of the respiratory system. This tropism, a location-specific expression, has been suggested as the reason why MERS can jump easily from camels to humans, but can’t jump easily between humans. Viral studies like the influenza one above further support this conclusion, since transmission appears to come primarily from upper respiratory tract infections.

Studies of influenza patients also suggest that individuals release drastically different amounts of virus into the environment. In one study, conducted at the Sendai Medical Center, authors found that less than half the patients (41%) expelled detectable amounts of viral gene copies, and only 3/56 samples contained viable viruses.  

Another study, this one conducted in North Carolina, recapitulated the same finding, that only about 40% of infected patients emitted detectable levels of virus into the air. However, they went one step further and compared the amounts of virus emitted by the different patients and found huge variability - some patents were releasing 32x more virus than others. 

Do they or Don’t they?

The data that has been collected about SARS-CoV-2 is far more limited in scope than the data that’s been collected on influenza A and other seasonal respiratory viruses. Information about SARS-CoV-2 aerosolization is severely limited, but the data we do have suggests a complex, multifaceted mode of transmission that will take a long time to fully understand.

What can be extrapolated from the limited data that we have available to us is that the location and progression of infection is fundamental to whether or not an individual sheds viruses into the air, and the characteristics of the air in the room are what decide if the droplets containing the virus remain aerosolized for enough time to infect someone else.

For this reason the six feet of guidance given for social isolation is our best bet, and it’s for this reason that the WHO, the CDC, and even the US Surgeon General have felt comfortable issuing no-mask advisories. 

What is also apparent, though, is that these advisories are motivated by a desire to drive down demand for masks that are in short supply and unambiguously necessary for frontline healthcare workers that are certainly exposed to high viral loads, especially when intubating patients. It’s a bizarre turn of events that, instead of recommending more DIY approaches to mask making, that these agencies instead attempted to mislead the public about the potential benefits. 

Wearing masks, even those made from simple materials that can be found around the house, is an effective approach to decreasing community spread. The WHO and CDC would have engendered much more goodwill by educating users about alternative solutions, rather than suggesting that masks are unnecessary. The demand for industrially produced masks certainly outstrips demand - but all of us can do our part by making our own.