The secret behind the mask

Abstract

Who should wear a mask? The symptomatic or the susceptible? The sick one or the healthy? We might ask these ‘simple’ questions whenever we encounter someone with the flu. Since the 2003 SARS epidemic, surgical masks have become an everyday accessory for many men and women in Asia. I frequently see people wearing masks on a train or bus in Hong Kong, where I live. When I refer to a mask here, I mean the disposable surgical masks that doctors or nurses often wear in hospitals. What does a surgical mask do? First, it filters out most of the large droplets or particles (those above 20 μm in diameter) and also some of the finer droplet nuclei (about 2–92% of those smaller than 20 μm) (Loeb et al., 2009). Second, it stops the exhalation puff of the wearer from being directly injected into air, instead redirecting it into the body’s thermal plume. This second characteristic is an attribute not recognized by many. So we have a quick answer. For a respiratory disease that is significantly transmitted by the large-droplet route, both the susceptible and infected can usefully wear a mask. For the susceptible, wearing a mask can reduce the risk of inhaling large droplets, and for the infected, wearing a mask can keep most of the large droplets behind the mask, and also ‘kill’ the exhalation jet. When the infected is small part of the population, as in a non-hospital environment, only infected individuals need to wear masks. For a disease that is transmitted by the airborne route via fine droplet nuclei, a susceptible person may not significantly reduce his or her infection risk by wearing a mask, but the infected can still reduce somewhat the risk of infecting others. What can a susceptible wear when someone who is infected with an airborne-transmitted disease is nearby? We can wear a mask that can filter out most (at least 95%) of the fine droplet nuclei, the so-called N95 mask. So a key question of importance for public health protection is whether any particular disease is transmitted by the airborne route via droplet nuclei, by the large droplet route, or by both routes. If both routes exist, what is their relative importance? How about the contact route? Unfortunately, we do yet know the answers to these questions for such a common and important disease as influenza. During the 2009 H1N1 influenza pandemic, lack of knowledge about routes of disease transmission led to contradictory recommendations from leading authorities regarding mask use by health-care workers. For example, WHO (2009) recommended the use of surgical masks, while the US National Academy of Sciences (IOM, 2009) recommended N95 masks. Such contradictions reveal the importance of studying the relative importance of each possible transmission route. Influenza is not alone in the study needs. Lower respiratory infections and tuberculosis are and will remain among the top seven causes of global disease burden in the near future. Here are two interesting observations when studying the literature of transmission routes of influenza. First, mathematical modeling results and empirical experience in the medical community do not agree. Infection of influenza and other respiratory diseases is observed to occur mostly in close proximity to the index patient. Such observations are often used as evidence for large droplet and close contact transmission. An observed transmission of further than 1–2 m is often cited as evidence for airborne transmission. Why has the 1–2 m distance become the boundary between droplet and airborne transmission? One explanation is that most large droplets would fall out of the exhalation jet within 1–2 m. However, using a modeling approach, Atkinson and Wein (2008) suggested that aerosol transmission is the dominant mode for influenza caused by rarity of close, unprotected and horizontally directed sneezes. Nicas and Jones (2009) also used a modeling approach to show that aerosol transmission is as important as hand contact and close contact. We need research to address and resolve this key disagreement. While there is less debate on the importance of airborne transmission routes for TB, chickenpox, and measles, controversies persist for influenza, the common cold, and SARS. Even reviewing the same literature, different authors draw opposite conclusions with regard to the quality of evidence for airborne infection, e.g. Tellier (2006) and Brankston et al. (2007) on influenza. Second, a line of reasoning suggests that because influenza infection mostly occurs in close proximity to the index patient, it is unlikely to be a result of airborne transmission via inhalation of droplet nuclei (Brankston, 2007). I can see a logical problem with this reasoning for droplet transmission. Although spray infection must occur in close proximity to the index patient, it is illogical to conclude that infection that occurs in close proximity must always be via droplet spray infection. In the indoor air science literature, we know that there is a substantial increase of airborne exposure of a susceptible person to droplet nuclei exhaled by the source patient when the two are in close proximity (within 1–2 m). This proximity effect suggests the possibility of short-range airborne transmission via droplet nuclei, in addition to the large-droplet transmission as making a contribution to close proximity infection. Nicas and Jones (2009) also have argued that close contact both permits droplet spray exposure and maximizes inhalation exposure to droplet nuclei and inspirable droplets. Hence for airborne transmission, we may need to consider both short-range and long-range airborne routes. We may also hypothesize that infection occurring in close proximity to the index patient includes both droplet-borne and short-range airborne routes. If this is true, then the short-range airborne infection pathway has probably been incorrectly grouped into the category of droplet infection ever since the pioneering work of Wells (1934). The mechanisms of droplet-borne and short-range airborne routes are different; hence, the effectiveness of alternative control methods is also different. For short-range routes, whether droplet or airborne via droplet nuclei, traditional methods such as ventilation dilution are not effective. However, there may be a role for personalized ventilation if droplet nuclei are important for short-range routes. During 2003, I had an opportunity to study the transmission route in the Amoy Gardens SARS outbreak in Hong Kong in which more than 300 people were infected (Yu et al., 2004). Roy and Milton (2004) wrote that there was a fitting symmetry between our study and John Snow’s investigation of a cholera epidemic 150 years ago. ‘…Snow’s independent investigation tested the hypothesis that cholera was waterborne. The official investigation …, however, concluded that transmission in the epidemic was airborne, caused by nocturnal vapors emanating from the Thames River – a conclusion that was consistent with the dominant paradigm of the time. Today, the situation is reversed. … they used computational fluid-dynamics and multizone modeling to test a hypothesis that the outbreak of SARS at the Amoy Gardens … was caused by airborne transmission. In the official investigation, airborne transmission was not seriously considered, because the current paradigm, as initially described by Charles Chapin in 1910, supports the belief that most communicable respiratory infections are transmitted by means of large droplets over short distances or through contact with contaminated surfaces.’ Snow’s findings were a serious blow to the miasma theory at the time. Subsequently, C Flügge showed, in 1897, that droplets from the nose and mouth contained bacteria, but did not travel more than 2 m. Flügge ‘concluded that true airborne infection other than within a few feet of the ‘infector’ was unimportant’ (Perkins, 1945). My medical collaborators introduced me the classical text by the eminent sanitarian Charles V Chapin (1910), which addressed the importance of contact infection. Following the work of Flügge, Chapin wrote: ‘most diseases are not likely to be dust-borne, and they are spray-borne only for two or three feet, a phenomenon which, after all, resembles contact infection more than it does aerial infection as ordinarily understood.’‘In reviewing the subject of air infection it becomes evident that our knowledge is still far too scanty, and that the available evidence is far from conclusive… It is impossible, as I know from experience, to teach people to avoid contact infection while they are firmly convinced that the air is the chief vehicle of infection….’ Here, Chapin deliberately emphasized the importance of contact transmission and also included spray-borne transmission as a type of contact infection. His view went unchallenged until the research findings a quarter century later on the significance of droplet nuclei by Wells. Chapin’s view probably still prevails today in the medical and public health communities. In his now classical study of airborne transmission, Wells (1934) examined the evaporation of falling droplets (apparently inspired by a falling raindrop study). He found that under normal room air conditions, droplets smaller than 100 μm in diameter would totally dry out before falling 2 m to the ground. This simple and elegant finding allowed Wells to establish the first scientific theory of droplets and droplet nuclei transmission based on the size of the infectious droplet at the origin of generation, i.e. mouth. According to Wells, droplet infection is transmitted by droplets larger than 100 μm in diameter. These large droplets rapidly settle out of the air by gravity, and so the infective range is limited to a short distance from the source. Conversely, airborne infection is attributed to dried-out infectious droplet nuclei that originate as droplets smaller than 100 μm as emitted at the mouth. These droplets would evaporate quickly, become droplet nuclei that remain suspended in the air for a longer time and could also be carried over significant distances by airflow prior to contact by a susceptible person. The study of Wells (1934) adds another important dimension to our indoor air science community. William Firth Wells, an engineer, working then at Harvard School of Public Health, was the first to show how the physics of droplets can be used to help explain disease transmission. His work makes clear that the indoor air community has an important role to play in regarding infectious diseases, studying the effects of room airflow, humidity and temperature, evaporation and dispersion of droplets, air cleaning, etc. That we can now distinguish between two transmission routes based on the droplet size allows us to consider different intervention methods such as ventilation or use of masks. Can the indoor air science community further contribute to understanding the transmission routes of respiratory infection, in particular for diseases such as influenza? The answer must be yes. A multidisciplinary approach is needed. Human beings have managed to eradicate smallpox, but around the world, the common cold is still all too common, and influenza still spreads in seasonal epidemics that cause true human suffering. Who is really behind the mask? We know a lot, but there are still too many unknowns. We have a role to play. Thanks go to Professor William W. Nazaroff for his useful comments on the first draft of this editorial.