Dataset: 11.1K articles from the COVID-19 Open Research Dataset (PMC Open Access subset)
All articles are made available under a Creative Commons or similar license. Specific licensing information for individual articles can be found in the PMC source and CORD-19 metadata.
More datasets: Wikipedia | CORD-19
Deep Learning Technology: Sebastian Arnold, Betty van Aken, Paul Grundmann, Felix A. Gers and Alexander Löser. Learning Contextualized Document Representations for Healthcare Answer Retrieval. The Web Conference 2020 (WWW'20)
Funded by The Federal Ministry for Economic Affairs and Energy; Grant: 01MD19013D, Smart-MD Project, Digital Technologies
Exposure to airborne pathogens is a common denominator of all human life. With the improvement of research methods for studying airborne pathogens has come evidence indicating that microorganisms (e.g., viruses, bacteria, and fungal spores) from an infectious source may disperse over very great distances by air currents and ultimately be inhaled, ingested, or come into contact with individuals who have had no contact with the infectious source [2–5]. Airborne pathogens present a unique challenge in infectious disease and infection control, for a small percentage of infectious individuals appear to be responsible for disseminating the majority of infectious particles. This paper begins by reviewing the crucial elements of aerobiology and physics that allow infectious particles to be transmitted via airborne and droplet means. Building on the basics of aerobiology, we then explore the common origins of droplet and airborne infections, as these are factors critical to understanding the epidemiology of diverse airborne pathogens. We then discuss several environmental considerations that influence the airborne transmission of disease, for these greatly impact particular environments in which airborne pathogens are commonly believed to be problematic. Finally, we discuss airborne pathogens in the context of several specific examples: healthcare facilities, office buildings, and travel and leisure settings (e.g., commercial airplanes, cruise ships, and hotels).
In summary, despite the various mechanistic arguments about which organisms can be potentially airborne and therefore aerosol-transmissible, ultimately, the main deciding factor appears to be how many studies using various differing approaches: empirical (clinical, epidemiological), and/or experimental (e.g. using animal models), and/or mechanistic (using airflow tracers and air-sampling) methods, reach the same consensus opinion. Over time, the scientific community will eventually form an impression of the predominant transmission route for that specific agent, even if the conclusion is one of mixed transmission routes, with different routes predominating depending on the specific situations. This is the case for influenza viruses, and is likely the most realistic.
Some bacterial and viral infections that have more than one mode of transmission are also anisotropic, like anthrax, plague, tularemia and smallpox: the severity of the disease varies depending on the mode of transmission [37, 89]. Older experimental infection experiments on volunteers suggest that this is the case for influenza, with transmission by aerosols being associated with a more severe illness [14, 90], and some more recent field observations are consistent with this concept. For anisotropic agents, even if a mode of transmission (e.g. aerosols) accounts for only a minority of cases, interruption of that route of transmission may be required if it accounts for the most severe cases.
Close contact is usually defined as being within 3 ft (~1 m) of the infector. When two or more students talk, small and large droplets will be sprayed from the infected student’s mouth. Large droplets are likely deposited on the mucous membranes of susceptible students, some small droplets are directly inhaled by the students who are talking with the infected student and some of the remaining small droplets are inhaled by others via the long-range airborne route. In this study, a close contact was counted when a face to face contact occurs between any two students who are within 3 ft (~1m). The quantity of virus (TCID50) changes via inhalation and deposition on mucous membranes according to factors such as relative height, distance and direction in which the two students were facing and the room’s airflow. Here we assumed that 50% of small droplets are inhaled by the student who is talking with the infected student and 30% of large droplets are deposited on that student’s face (10% of droplets on the face are deposited directly on the mucous membranes). Other small droplets diffuse into the air, and large droplets are deposited on surfaces nearby.
We observed 3526 close contacts between students in an office over the course of 5 days. Figure 2 shows the association between the duration of close contact and percentage of contacts. The frequency of close contact per person is 9.64 h−1 per student, including active and passive contacts. Each student spent an average of 9.86% of their time in close contact. The average duration and the mean duration of close contact were 53.8 s and 17 s, respectively.
This study is the first to examine the long-term variation of the distribution of airborne adenoviral particles and Mycoplasma pneumoniae in the pediatric emergency room and pediatric outpatient department of the Children's Hospital in Taiwan. The detection rate of airborne adenovirus DNA products in the pediatric emergency room clearly exceeded those of Mycoplasma pneumoniae DNA products. Airborne Mycoplasma pneumoniae DNA products were detected in around 46% of samples from the pediatric outpatient department of the hospital. The distribution of microbial species in the pediatric public areas warrants further attention from the department of environmental safety and health to reduce the risk of microbial exposure in humans. The possibility of adenovirus and Mycoplasma pneumoniae that transmitted by the airborne route also warrants further investigation.
In this study, the detection rates of airborne adenovirus DNA products peaked in summer in the pediatric outpatient department and the pediatric emergency room, respectively. Furthermore, airborne Mycoplasma pneumoniae were frequently detected in the pediatric outpatient department and the detection rates were highest in autumn and winter seasons. This results suggest that the ventilation rate in pediatric public areas (including the emergency room and outpatient department) of hospitals should be increased during the peak seasons of adenovirus and Mycoplasma pneumoniae contamination to reduce the airborne bioaerosol levels. The limitation of this study was lack of environmental characteristics measurement along with airborne adenovirus and Mycoplasma pneumoniae sampling. Further study will enable the relationship between indoor environmental characteristics (including temperature, relative humidity and CO2 concentration) and the airborne adenovirus/Mycoplasma pneumoniae to be elucidated.
The air sampling time in the pediatric outpatient department was shorter than that in the pediatric emergency room. The main reason was that the pediatric outpatient department of the hospital was close at nighttime. The monthly detection rate of airborne adenovirus DNA products in the filter samples in the pediatric outpatient department was lower than that in the pediatric emergency room, but the measured viral concentrations in March, May, June, and July in the pediatric outpatient department considerably exceeded those measured in the pediatric emergency room. Additionally, Mycoplasma pneumoniae exhibited not only a higher detection rate but also higher concentrations in the pediatric outpatient department than in the pediatric emergency room. The possible cause may be the indoor unfair ventilation and more patients with Mycoplasma pneumoniae infection in the pediatric outpatient department. The results demonstrate that the mean CO2 concentration in the pediatric outpatient department exceeded that in the pediatric emergency room. Therefore, the ventilation system should be adjusted periodically and the air change rate should be increased during peak seasons with high viral concentrations in the pediatric public areas. Also, high-efficiency particulate air (HEPA) filters should perhaps be installed in the ventilation systems in the pediatric public areas of hospitals to maintain indoor air quality. In addition, the results indicated that monthly distributions of detection rates of adenovirus and Mycoplasma pneumoniae in the air samples were inconsistent with the median concentration distributions of airborne adenovirus and Mycoplasma pneumoniae in the pediatric public areas. The possible reason might be included: 1) the air samples in the pediatric public areas were only collected twice in a week; 2) the exact number of infected patients in the pediatric public areas was unknown.
A previous investigation found that the number of patients with influenza was closely correlated with the number of air samples that contained influenza A RNA (r = 0.77). Tseng et al. study indicated that the influenza A virus concentrations in the air were significantly correlated with the number of patients with low respiratory tract infections (r = 0.43, p = 0.02), but no correlation occurred between the viral concentrations and the number of patients with upper respiratory tract infections (r = 0.33, p = 0.06). During the study period, there was not outbreak occurrence of adenovirus and Mycoplasma pneumoniae infections in the hospital. There were a great number of patients and visitors in the hospital, and not all of patients with respiratory tract infection were tested for adenovirus or Mycoplasma pneumoniae infections, so it would be underestimated the number of patients with adenovirus or Mycoplasma pneumoniae infection in the pediatric public areas. The relationship between the number of infected patients and the number of air samples with detected certain pathogens warrants further investigation.
Indoor air quality is important to prevent infection and thereby protect patients and health care workers in hospitals. The quality of indoor air depends on heating ventilation and air conditioning (HVAC) systems, the number of persons, human activity and other factors (i.e., the outdoor air quality),. Until recently, no standard method had been established for detecting specific viral aerosol particles. Tolerable levels of airborne viruses in hospitals have not yet been established. Airborne particle concentration has been suggested to be an indicator of microbial contamination. In this study, PM10 and PM1 levels were higher in the pediatric outpatient department than in the pediatric emergency room. Hence, further investigations must be performed to characterize the relationship between particulate matter levels, such as PM10 and PM1, and the concentrations of airborne viruses and bacteria.
Westwood et al. (1966) demonstrated that inhalation of a single PFU of a submicrometer vaccinia aerosol was sufficient to infect rabbits. Airborne rabbit pox was similarly infectious. They demonstrated rabbit-to-rabbit airborne transmission of rabbit pox in each of seven trials by placing uninfected rabbits in separate cages in the same room with infected animals. They also infected rhesus monkeys using submicrometer aerosols of variola.
In one of the earliest extensive animal models of smallpox, Brinckerhoff and Tyzzer (1906) reported the effect of inoculating cynomologus monkeys with variola at different sites. Inoculation of mucus membranes of the lip, palate, and nose produced local lesions, but generalized rash occurred in only 10% of animals. Inoculation through the skin produced a local lesion and a generalized eruption in 70–80% of animals. Animals inoculated by scratching the tracheal mucosa through a rigid bronchoscope all developed a generalized rash, and one developed a variolous bronchitis and pneumonia. Laryngeal instillation of dry pustule contents produced infections while instillation of powdered crusts did not. Inhalation exposures to an atomizer spray of vesicle contents infected only one of five monkeys; however, the particle size distribution and type of atomizer were not reported.
Hahon and Wilson demonstrated that infection of Macaca irus with high dose [5 × 105 PFU] fine particle (<5 μm) variola aerosols produced a disease that simulated human smallpox (Hahon and Wilson, 1960; Hahon, 1961). The initial site of virus replication was the lung, with subsequent appearance of virus in the nasopharynx and nares. Peak concentrations of virus per gram of tissue were higher in the lung than in the upper respiratory tract; the peak in lung tissue occurred during the incubation period and lung levels declined during the secondary viremia and exanthem. Whether the time course and viral concentrations in lung in this animal model produced by inhalation of high dose aerosols mimicked that in humans with natural infection is doubtful. However, it may be relevant to the first generation of cases exposed to concentrated aerosols in a biological attack. In a relatively recent experiment, (Kalter et al., 1979) a female chimpanzee became infected with variola while housed in the same room, but without direct contact, with two infected chimpanzees. She developed a generalized rash and was reported to have had more severe constitutional symptoms than the other chimpanzees infected by dermal inoculation or direct contact. The authors concluded that she was infected via aerosol.
The animal data show that artificial respirable aerosols were effective means of producing poxvirus infections, that the infectious dose by the airborne route could be very low, and that animal-to-animal airborne transmission of rabbitpox and variola was observed. They also suggest that inoculation of mucus membranes was less effective at producing a generalized rash than was exposure of the lower respiratory tract.
Tuberculosis is a localized or systemic, but most often respiratory bacterial illness caused by mycobacteria belonging to the Mycobacterium tuberculosis complex.
For tuberculosis (TB), definitive experimental evidence of airborne transmission being necessary and sufficient to cause disease was provided in a series of guinea-pig experiments [30, 31], which has been repeated more recently in a slightly different clinical context. Numerous other outbreak reports have confirmed the transmissibility of TB via the airborne route [33–35], and interventions specifically targeting the airborne transmission route have proven effective in reducing TB transmission.
Several infectious diseases are known to be transmissible through the airborne route, such as tuberculosis and measles. The only known disease transmitted only through the airborne route is tuberculosis, as reported by Roy and Milton. Aerosol transmission of other diseases could be preferential or opportunistic. Therefore, it is difficult to assess the importance (or not) of the airborne route in disease transmission. The mode of transmission of some diseases is ambiguous. The evidence of severe acute respiratory syndrome (SARS) airborne transmission was first assessed by indirect evidences such as modeling and epidemiological studies. Moreover, the possible airborne transmission of non-respiratory diseases, like Norovirus, is a subject of investigation.
For many diseases, dissociating transmission routes such as indirect contact, exposure to large droplets and aerosol transmission through aerosols can be complicated, even in controlled laboratory environments. The World Health Organization considers disease transmission with particles >5 µm as droplets transmission and with particles <5 µm as aerosols transmission.
The size of the particles involved in the natural transmission of diseases through the airborne route is hard to establish, especially for viral diseases. In fact, only a few studies have looked at the particle size of airborne viruses that can be found in the environment. Anderson 6 stage cascade impactors, National Institute for Occupational Safety and Health (NIOSH) two-stage bioaerosols cyclone samplers, and Sioutas personal cascade impactors have been used in agricultural and hospital settings. In all these studies, viruses were found in all air sample stages, meaning that large particles as well as small particles can carry viruses. More recently, in a laboratory setting, experiments using ferrets and particle impactors of various cut-off sizes demonstrated that influenza virus can be transmitted via droplets (15.3–5 µm) as well as airborne particles (5–1.5 µm).
Information on the infectious state of airborne viruses is sparse. Culture on appropriate cell lines is still the gold standard to assess virus infectivity. However, the culture of airborne viruses faces several challenges: (1) low concentrations of viruses in the air require large air volume sampling to allow detection (meaning extensive air sampling periods or the use of high-flow air samplers); (2) viruses can be damaged during air sampling; (3) environmental contaminants can interfere with virus or host cell growth (bacteria, mold, dust, etc.).
The use of animals in laboratory settings can overcome most of these challenges. The virus source can be a sick human, an infected animal, or an artificially generated aerosol. By exposing animals to airborne viruses, air sampling can be avoided (preventing virus damage) as well as laboratory virus culture bias in detection. As an example, using animals instead of air samplers can lead to the demonstration that airborne viruses can or cannot infect healthy animals and also that airborne viruses can remain (or not) infectious long enough to travel to a new host. Using a sick animal or human as an aerosol source has also demonstrated that a sick subject can emit aerosols that can potentially infect other susceptible hosts.
Unfortunately, the exposure of healthy animals to aerosols emitted by another animal over several days cannot be performed in commercially available apparatus settings. Indeed, cages designed for animal aerosol exposure are meant for a few minutes per day exposure and cannot be used for housing animals for several days. In contrast, animal cages designed to house subjects over extended periods of time are not airtight, and therefore provide limited information about airborne transmission. These cages can be used to prevent direct contact between index and healthy animals and can be placed at various distances but cannot control the size of particles traveling between cages.
In this study, we designed, constructed and tested a system composed of three airtight cages to study the transmission of infectious agents between animals through large droplets and through airborne particles. The system can house three ferrets for up to 10–12 days and is designed to prevent direct contact between animals. We designed a particle separator to prevent large droplets transmission between cages. The cage system is under negative pressure, with high-efficiency particulate air (HEPA) filters on the air inlet and outlet for the users’ and environment’s protection. This communication describes the main components of this cage system, the particle separator validation using standard aerosol generators as well as a test trial with ferrets and the influenza virus.
Students move, make contact with other students, and touch surfaces in the office. Influenza A virus will be transferred between the hands and surfaces over time. From Figure 3a, the virus on the hands of the infected student increases rapidly and reaches a balance because limited numbers of his or her private surfaces share the virus. The hands of other students (susceptible) will be gradually contaminated, and virus on the hands of susceptible students almost reach a balance after 3 h (12 a.m.). The private surfaces of infected students are highly contaminated, and the quantity of virus (TCID50) on the private surfaces of the infected student is almost three orders of magnitude of that on the private surfaces of susceptible students. Public surfaces are dirtier than the private surfaces of susceptible students, and they also play important roles in the spread of infection like hubs in the surface touch network. The cumulative quantity of virus (TCID50) on the mucous membranes of susceptible students expresses each student’s intranasal dose and gradually increases over time.
The quantity of virus (TCID50) on the floor is relatively low because we assumed that most large droplets generated by talking, coughing and sneezing are deposited on the top surfaces of the desk, if the infected student is sitting in his or her own seat. In a 1-day simulation, we found that the respiratory dose per day of each susceptible student from the long-range and short-range airborne routes are 0.25 and 0.17 TCID50, respectively, and the intranasal dose per day from the fomite and droplet spray routes are 63.81 and 185.24 TCID50 (Figure 3b). All results are average values from 1000 simulations. Based upon the dose-response parameters from two routes, the total infection risk for each susceptible student during 1 day in the office is 8.75%, of which 54.31%, 4.23% and 44.46% are contributed by the long-range airborne, fomite and close contact routes. The class coordinator usually has more frequent interaction with other students, and we assumed that the probability of the class coordinator touching others’ desks and chairs and talking with others is twice that of the other students. The infection risk of the monitor is 13.79% (Figure 3b). The virus distribution on the surfaces of the desks and chairs of the class coordinator and the other students is shown in Figure S1 (Supplementary Materials). The class coordinator has a higher infection risk than the other students. There are 57 types of sub-surfaces, and the final quantity of virus (TCID50) on each type of sub-surface (TCID50 per surface) is shown in Figure 3c. The quantity of virus (TCID50) is much higher on the private surfaces around the infected student (approaching 800 times) than around susceptible students. Keyboards, headphones, desktops, mice and mobile phones are the five most-contaminated private surfaces around the infected student. The top of the seat back, the right chair arm, the desktop, the left chair arm and the top of the left desk’s fence are the five most-contaminated private surfaces around the susceptible students. Air conditioning (AC) controllers, printer touch screens, cabinet handles, tissue dispensers and the printer drawer are the dirtiest of all public surfaces. Surfaces with small areas, such as headphones and the buttons on the AC controller usually have a high virus concentration.
If natural smallpox was initiated through the upper respiratory mucosa, then an early asymptomatic mucosal infection would be expected. To investigate this, Sarkar and colleagues performed pharyngeal swab surveys of household contacts (Sarkar et al., 1973a, 1974) 4–8 days following onset of rash in the index cases. They found that contacts with positive throat cultures often did not develop smallpox. In one survey, (Sarkar et al., 1973a) 10% (Westwood et al., 1966) of 328 contacts had positive swabs, but only 12% (Kaplan et al., 2002) of those with positive swabs developed smallpox. Among 59 unvaccinated contacts 27% (Miller, 1957) were culture positive, but only one developed smallpox. All subjects were vaccinated at the time of examination. However, vaccination four or more days after exposure is usually considered to be too late to prevent disease. The observation that disease did not develop in 94% of persons with mucosal infection suggests that, even in unvaccinated contacts, mucosal infection may not have been sufficient to initiate disease.
Sarkar and colleagues also showed that the oropharyngeal excretion of virus was greatest during the first days after the rash erupted and generally resolved at most 2 weeks following onset of rash (Sarkar et al., 1973b). Rao et al. found that oropharyngeal excretion was greatest in the most severe, hemorrhagic cases and corresponded with the period of infectiousness (Rao et al., 1968). In contrast to oropharyngeal excretion, scabs contained large quantities of virus regardless of disease severity (Mitra et al., 1974) and were shed for another week or more after throat cultures were negative. Scabs alone, however, were not associated with further cases (Rao et al., 1968; Mitra et al., 1974).
The apparent lack of infectiousness of scab associated virus has been attributed to encapsulation with inspissated pus (Fenner et al., 1988). Henderson's theory about the importance of small particles may provide a straightforward mechanism for why encapsulated virus, simply by entrapment in large particles, had low infectious potential.
Sarkar et al. (1973a) were concerned that asymptomatic contacts could have been infectious because their throat swab viral titers were similar to those of milder smallpox cases. A paradox arose from these data because there was never evidence of infection arising from asymptomatic household contacts. Yet, oropharyngeal secretions were thought to be the primary source of infectious virus particles. An explanation may be that oropharyngeal excretion of virus was merely temporally correlated with excretion of virus from elsewhere in the respiratory tract and not the actual source of fine particles virus aerosols.
The large spray of particles from sneezing visualized by high speed photography consists of particles down to about 10 μm in diameter (Papineni and Rosenthal, 1997). Smaller particles may also be dislodged from the upper airways by the turbulence of sneezing, coughing, and talking, but will mostly be larger than 2.5 μm in diameter. Recent studies, however, show that the healthy lung generates abundant fine particles (100–1000/l with size <0.3 μm diameter) during normal breathing (Fairchild and Stampfer, 1987) that do not arise from the oropharynx; condensates of these particles are the subject of recent reviews (Mutlu et al., 2001; Hunt, 2002). Such particles could carry variola virus (0.2–0.3 μm diameter), would remain airborne in indoor air for many hours, and would be deposited primarily in the lower airways after inhalation.
There is some evidence that variola was present in the lung and potentially available for aerosolization. Animals infected by inhalation produced high concentrations of variola in the lung (Hahon and Wilson, 1960). Fenner et al. (1988) regarded bronchitis and pneumonitis as a part of the normal smallpox syndrome, especially in the more severe cases which were also the most infectious, (Rao et al., 1968) although specific lesions were less frequent in the lower trachea and bronchi. Systematic evaluations of viral excretion in the lower respiratory tract of non-fatal cases were not reported. Thus, if some degree of pneumonitis with pulmonary excretion of virus and exhalation of fine particle variola aerosols was a feature of clinical smallpox but was not a feature asymptomatic household contact with positive throat cultures, then the paradox would be resolved.
During the study period, the air temperatures and relative humidity were 16.1–26.6°C and 42.5–77.7%, respectively, in the pediatric emergency room, and 13.7–24.8°C and 50.4–84.7% in the pediatric outpatient department. The CO2 concentrations were 404–704 ppm in the pediatric emergency room and 372–2,575 ppm in the pediatric outpatient department. The mean concentrations of ≦10 µm (PM10) and ≦1 µm (PM1) particles were 8.9 µg/m3 and 17.0 µg/m3, respectively, in the pediatric emergency room, and approximately 105 µg/m3 and 167 µg/m3, respectively, in the pediatric outpatient department.
Filter samples without airflow passage were selected as negative controls. The adenovirus and Mycoplasma pneumoniae were not found in the blank filters. Also, PCR positive rates of the positive controls containing adenovirus or Mycoplasma pneumoniae in the sampling filters are 100%. The monthly detection rate (18.3%) of the adenovirus in the air samples of the pediatric emergency room of the medical center was clearly higher than that of Mycoplasma pneumoniae (0.9% of the air samples) (p = 0.009) (Fig. 1). Around 45.8% of the air samples from the pediatric outpatient department contained Mycoplasma pneumoniae. Other pathogen, adenovirus (9.8% of the air samples), was also found in the pediatric outpatient department.
In this study, airborne particles that contained adenovirus were detected in the pediatric emergency room throughout the year except in March to May, and October (Fig. 2A). Detection rates of airborne adenovirus DNA products in the filter samples in the pediatric emergency room were high in July (33.3%) and August (66.7%). The airborne Mycoplasma pneumoniae DNA products were detected in the pediatric emergency room only in July (11.1% of the air samples). The airborne particles that contained adenovirus concentrations ranged between <10 copies/m3 and 104 copies/m3 in the pediatric emergency room. The median concentration of airborne adenovirus DNA was highest in December. Only one air sample in the pediatric emergency room was positive for Mycoplasma pneumoniae DNA product (645 copies/m3).
The detection rate of adenovirus-containing particles in the air samples in the pediatric outpatient department peaked in July (66.7%) (Fig. 2B). However, airborne particles that contained Mycoplasma pneumoniae were found in the pediatric outpatient department throughout the year except in February to June. High detection rates of airborne Mycoplasma pneumoniae DNA products in the filter samples were found in August (83.3%), September (100%), October (88.9%), and November (100%), and January (100%). The detected concentrations of adenovirus in the air in the pediatric outpatient department ranged from 48.4 copies/m3 to 461 copies/m3, and the median concentration was highest in March. The concentrations of Mycoplasma pneumoniae in airborne particles ranged between 114 copies/m3 and 9.9×104 copies/m3 in the pediatric outpatient department. The median concentration of Mycoplasma pneumoniae DNA product in the air of the pediatric outpatient department was highest in October.
While it is unclear exactly how MERS is contracted, it is likely to spread via an infected person’s respiratory secretions like other coronaviruses. To date, there has not been widespread sustained community human-to-human transmission. It appears that close contact with an infected person is necessary for disease transmission. Close contact is defined as encountering a patient without appropriate protective gear within six feet or being in a care room for prolonged periods or having direct exposure to infected secretions. Healthcare facilities have reported spread from person-to-person much more so than in communities, possibly when suboptimal infection control was practiced for patients with higher viral loads than those not hospitalized.7
Reported cases have been linked to countries in and near the Arabian Peninsula either for persons who live in, have traveled to, or have had contact with an infected person who had been in the region. MERS is a zoonotic virus that is transmitted from animals to humans. It is believed to have originated in bats and then to have been transmitted to camels sometime in the distant past. According to epidemiologic and surveillance data, there is a strong likelihood that dromedary (one-hump) camels (Figure 2) serve as a reservoir for zoonotic transmission of the virus to humans.8 This has resulted in warnings to avoid close contact with camels and not drink raw camel milk or urine, or ingest raw camel meat.
The origins of infections resulting from droplet and airborne transmission are at the intersection of the clinical manifestation of disease, the site of infection, the presence of a pathogen, and the type of pathogen. Thus, when investigating the origins of droplet and airborne infections, there are several well-known primary sources of infectious particles (see Table 1): vomiting, toilet flushing (i.e., toilet water aerosolization), sneezing, coughing, and talking. Moreover, toilet bowls, the water in them, and toilet seats may harbor infectious particles after the initial flush, making additional aerosolization of infectious particles possible with additional flushes for as long as 30 minutes after the initial flush. Particle desiccation, discussed above, is important in this context. A single sneeze, for example, generates as many as 40,000 large droplet particles; most will desiccate immediately into small, infectious droplet nuclei, with 80% of the particles being smaller than 100 μm.
The transmission of infectious diseases via airborne or droplet routes may also depend on the frequency of the initiating activity. For example, while a single sneeze may produce more total infectious particles than a cough [11, 28, 65, 66], Couch et al. reported that coughing is more frequent than sneezing during infection with Coxsackievirus A. This finding suggests that coughing is a more likely method of airborne transmission for this disease than sneezing. As coughing is also a common symptom of influenza infection [68, 69], it may also contribute to the airborne transmission of this pathogen.
Finally, infectious individuals are not always the immediate source of airborne infectious particles. Many people spend considerable time in office buildings, for example, and as a result become exposed to airborne pathogens that originate from nonhuman sources (e.g., molds, toxins produced by molds, pollen, pet dander, and pest droppings) [70–77]. The health effects associated with naturally occurring indoor biological air pollutants include disease, toxicoses, and hypersensitivity (i.e., allergic) diseases [70–77]. In addition, exposure to indoor biological air pollutants has been associated with “sick building syndrome,”a set of nonspecific symptoms that may includeupper-respiratory symptoms, headaches, fatigue, and rash and“appear to be linked to time spent in a building, but no specific illness or cause can be identified.”.
The emergence of new unknown virus in conjunction with high fatality rate of the disease has led to major public health and international concern. The WHO guidelines apply a relatively sensitive definition of suspected cases and emphasize the importance of a high index of clinical suspicion for diagnosis owing to the high mortality rate. Some countries, especially those in endemic regions, have developed their own preparedness and response plans which are based on WHO and Centers for Disease Control and Prevention (CDC)’s recommendations. The Oman’s Ministry of Health implemented a national plan which was based on strengthening five pillars of action, including: (1) public health surveillance and contact management. Field visits were conducted to every confirmed case by the national public health services and exposed individuals were monitored for 14 days after the last exposure; (2) building laboratory capacity, including diagnostic capacity with primers for MERS-CoV testing, and training laboratory personnel on triple-packing and shipment of samples. Furthermore, training to first responders and intensivists on how to collect nasopharyngeal samples; (3) infection prevention and control, including mask-fit testing for all healthcare workers who could be involved in patient care; (4) case management; and (5) risk communication. The government of the Republic of Korea summoned a Rapid Response Team following the outbreak in their country. The team was composed of infectious disease specialists and infection control professionals, and they established national guidelines for the diagnosis and management of MERS-CoV infection. Together with the epidemiology investigation team of the local government, control strategies were discussed, which included: (1) contact tracing; (2) surveillance of polymerase chain reaction testing of healthcare workers and patients according to their level of contact; (3) preemptive isolation of pneumonia cases; (4) environmental disinfection; and (5) cleaning and enforcing the use of personal protective equipment (PPE) among healthcare providers. The possibility of MERS-CoV occurring in Israel is high given Israel’s geographic location in the Middle East and the thousands of Israeli Moslems who make the pilgrimage to Mecca (the Hajj) each year. Therefore, the Israel Ministry of Health (IMOH) has drafted preparedness guidelines that generally follow the CDC guidelines, although the CDC did not include Israel among the countries at risk. These guidelines recommend laboratory evaluation for all healthcare workers with a severe acute respiratory illness of unknown etiology, and in cases of clusters of severe respiratory symptoms of known etiology. In addition, the approval of the public health services is required before any case may be designated a suspected MERS-CoV infection, thereby ensuring early involvement on a national level in every instance of the disease. Information regarding MERS-CoV was disseminated by the distribution of leaflets and placement of informative posters at Ben-Gurion International Airport and three land-border crossings between Israel and Jordan.
Crowdedness and outdoor air ventilation per person are important for the spread of airborne infectious diseases in rooms such as dorms where people spend a lot of time. Respiratory viruses can be transmitted through air so that transmission is modulated by outdoor air supply rates. Further studies are warranted.
Prevention of MERS-CoV transmission involves avoiding exposure. Travelers to regions where MERS has been detected should avoid close contact with potentially infected persons or dromedary camels. Healthcare personnel must practice strict standard, contact, and airborne precautions while caring for patients under investigation (including symptomatic close contacts) as well as patients with probable or confirmed MERS infections. Laboratory workers and others collecting and handling specimens for potential MERS patients should adhere to the same guidelines. Adequate respiratory protection is particularly important when performing aerosolizing procedures.
In summary, we conducted a 7-week pilot study to monitor for aerosolized respiratory viruses among inpatients in a pediatric ward in Singapore. We found molecular evidence of influenza A virus and adenovirus, demonstrating the potential for airborne transmission. To comprehend this potential risk of transmission, our proof-of-concept project might be expanded in the future to specifically study patients with known positive clinical infections (e.g. through nasopharyngeal swabs) and determine how far away viable viruses can be detected from a patient’s bedside. Additionally, future studies might involve more comprehensive demographic and clinical risk factor analyses to help us better understand phenomenon such as super-spreading. Lastly, bioaerosol surveillance might be useful in monitoring clinical populations for incursions of novel respiratory viruses, especially since aerosol sampling in shared clinical spaces often requires no informed consent.
Typical hospital air quality indices are temperature, relative humidity,1 and levels of carbon dioxide, particulate matter,2,3 toxic metals,3 volatile organic compounds,1 bacteria,2,4–6 fungi,5,7,8 and viruses.9 Previous studies of air quality primarily focused on intensive care units,4,7,10,11 operating rooms,12,13 negative-pressure patient isolation rooms,6,9 and public areas in hospitals.3,14–19 However, few studies assessed viral contamination in the patients’ rooms in a given hospital ward.
Airborne microorganisms in a hospital can infect susceptible patients.20 Aerosolized droplets are generally 4 to 8 μm in diameter,21 while most viruses are 25 to 400 nm in length.22 In nature, airborne viruses associate with larger particles and aggregate.23–25 However, the size distribution of airborne viral particles is rarely determined. Two mechanisms underlie the person-to-person transmission of viral infections of the respiratory system: exposure to large-droplet infectious nuclei that remain suspended in air for short periods, and exposure to small-particle infectious nuclei that can remain suspended in air for long periods.26 Generally, large-particle aerosols are believed to account for viral transmission.
Viral infections of the respiratory system are very common. In Taiwan, the predominant viruses isolated from patients with respiratory infections are enterovirus, respiratory syncytial virus (RSV), influenza A and B viruses, adenovirus, cytomegalovirus, herpes simplex virus-1, and parainfluenza virus.27 Enterovirus causes herpangina, hand-foot-and-mouth disease, myocarditis, encephalitis, and death. RSV is the most common pathogen of the lower respiratory tract in infants28 and a common cause of nosocomial infections in pediatric wards.29 Influenza A and B viruses cause seasonal epidemics in Taiwan, especially in winter.30 Adenovirus causes acute respiratory tract infections in children younger than 5 years of age and circulates throughout the year.31 Further, Mycoplasma pneumoniae, a small bacterium, is a common pathogenic agent of community-acquired pneumonia (CAP) in children and young adults.32,33
RSV has been detected in air samples collected at distances of 30 to 700 cm from the head of patients’ beds,34 and RNA analysis of such samples at distance of 700 cm from the bedside may be useful for identifying small RSV-containing particles in hospital wards. Moreover, ubiquitous objects (e.g., telephones, door knobs, tables, air filters, ventilators) in hospitals have been shown to harbor adenovirus,35,36 varicella-zoster virus,37,38Staphylococcus aureus, and Pseudomonas aeruginosa.39
Few studies have investigated the concentrations of respiratory viruses and M pneumoniae aerosols in the rooms of pediatric wards. To provide more detailed information, we measured the concentrations of 4 respiratory viruses common in children (enterovirus, influenza A virus, RSV, and adenovirus) and M pneumoniae in air samples collected at 2 locations (relative to the head of the bed) in patient-occupied rooms in the pediatric wards of a university-affiliated hospital in northern Taiwan. We also inspected the objects in the pediatric ward rooms to determine whether they were contaminated.
Figure 3 presents the fitness between the reported attack rates and predicted infection risks in different scenarios, while Table 1 lists the scenarios with the best fitness (the minimum RSS) for the seven types of serving pathways. According to the dose-response relationship model, the infection risk increased with the viral load, L0, and the dose-response parameter of the mucous membranes, ηm. Excessively small or large values of these two parameters yielded correspondingly too low or high infection risks that deviated considerably from the attack rates. Therefore, the best fitness was reached with moderate values of ηmL0, as shown in Figure 3 and Table 1.
Of the seven pathways, Pathway 1 exhibits the best fitness because the corresponding predicted infection risk distribution was qualitatively similar to that of the attack rates. As shown in Table 1, the infection risk was highest at Table 2 and decreased from Tables 1–6. As shown in Figure 3, the scenario with the best fitness is very similar to the baseline scenario, indicating that the index patient in this outbreak probably shed a viral load, L0, similar to those reported by the literature, with a ηm of 0.1415/genome copy.
Aerosol-generating medical procedures (AGMPs) are increasingly being recognized as important sources for nosocomial transmission of emerging viruses. Intubation was investigated as a possible cause of Ebola virus (EBOV) transmission among health-care workers (HCWs) in the United States. Additionally, the high rate of nosocomial transmission of Middle East respiratory syndrome coronavirus (MERS-CoV) and Severe Acute Respiratory Syndrome coronavirus (SARS-CoV) caused speculation about the role of AGMPs. Crimean–Congo hemorrhagic fever orthonairovirus (CCHFV), was also associated with nosocomial infection secondary to AGMPs. While guidelines were developed for performing AGMPs on patients with certain viral infections, assessing and understanding the risk that specific viruses and AGMPs pose for nosocomial transmission could improve infection control practices, as well as reveal relationships in virus transmission.
Despite the perceived importance of AGMPs in nosocomial transmission of viruses and other infectious agents, scarce empirical or quantitative evidence exists. In order to assess the risk that certain viruses and AGMPs create for nosocomial transmission, we first need to identify potential AGMPs and viruses. The second step is then to determine the risk associated with these viruses and procedures, either through retrospective analysis, investigating the circumstances of nosocomial transmission, or through experiments, such as using air sampling during AGMPs to determine the risk of generating infectious virus-laden aerosols. Lastly, we can use this knowledge to re-evaluate current guidelines and communicate which viruses and AGMPs pose the highest risk for nosocomial transmission.
Chang Gung Children's Hospital is a part of Chang Gung Memorial Hospital, which is a 3700-bed university-affiliated teaching hospital in northern Taiwan that provides primary to tertiary care. The general pediatric wards of the Children's Hospital currently contain 220 beds distributed among 6 floors, 2 of which house most of the patients with infectious diseases. In each pediatric room, there are 1 to 3 beds and, hanging on the wall near the door, 1 bottle of alcohol-based hand-sanitizing solution. There are also 3 one-bed and 2 two-bed negative-pressure isolation rooms. Patients with pulmonary tuberculosis, measles, or varicella are always placed in the negative-pressure isolation rooms.
In general practice, patients with clinically suspected enteroviral infections, such as herpangina or hand, foot, and mouth disease, are segregated in a given set of rooms. The care providers who stay in these rooms (e.g., parents, grandparents) receive specific instructions regarding infection control procedures, such as wearing a mask, maintaining hand hygiene, and using disinfectants. The caregivers of patients with documented influenza (a positive rapid test) also follow these guidelines as much as possible. There are no specific guidelines for the caregivers of patients with adenovirus, RSV, or M pneumoniae infection, although hygiene is emphasized. The air in the ward rooms was air-conditioned but not heated during the study period.
The study period was from October 2009 to September 2010. For this study, 75 hospitalized children who we suspected were infected with enterovirus, influenza A virus, RSV, adenovirus, or M pneumoniae were recruited after obtaining written informed consent from their parents or guardians. Ultimately, it included 58 patients with infections subsequently confirmed via virus culture or polymerase chain reaction (PCR) (enterovirus, 17 of 24 patients; influenza A virus, 13 of 16 patients; RSV, 13 of 15 patients; adenovirus, 9 of 13 patients; and M pneumoniae, 6 of 7 patients). Air and object surface samples were obtained from the rooms of these patients. The study protocol was approved by the institutional review board of the hospital (97-1394B).
Medical procedures that have the potential to create aerosols in addition to those that patients regularly form from breathing, coughing, sneezing, or talking are called AGMPs. While there are many suspected AGMPs, few AGMPs were confirmed to generate aerosols. In order to determine which AGMPs could be important for nosocomial virus transmission, we first need to characterize what aerosols are and how they are created.
Aerosols are particles suspended in air that can contain a variety of pathogens, including viruses, and there is ongoing debate about how to classify them. Many divide aerosols into the categories of small droplets (which some exclusively call aerosols) and large droplets, with small droplets having the potential to desiccate and form droplet nuclei that travel long distances, while large droplets do not evaporate before settling on surfaces. Classifying aerosols by their initial size is relevant in relation to their dispersal patterns, but it is also important to classify aerosols according to where they deposit in the respiratory tract because pathogenesis can be influenced by whether a virus deposits in the upper respiratory tract (URT) or lower respiratory tract (LRT). Dispersal and deposition depend on a variety of factors, and there is no exact cutoff for small and large droplets. Some authors use ≤5 µm in diameter as a cutoff for small droplets, while another possible cutoff between aerosol types is 20 µm, since aerosols ≤20 µm in diameter can desiccate to form droplet nuclei, and aerosols ≥20 µm do not deposit substantially in the LRT.
Often the term airborne transmission is used to describe infection by small droplet aerosols and droplet nuclei, while droplet transmission refers to the route of large droplet aerosols. Since aerosols can be of multiple sizes, we use the term aerosol transmission to generally describe transmission through the generation of infectious small and large droplet aerosols. In addition to these modes of transmission, AGMPs may also create opportunities for direct contact and fomite transmission, which may be difficult to distinguish.
HCWs are considered to be at risk for nosocomial virus transmission from both small and large droplet aerosols, for both seem to play a role in human-to-human virus transmission. Small droplets can be inhaled into the LRT, while large droplets can splash into the eyes or mouth and deposit in the URT. Certain respiratory viruses, like influenza A virus, are believed to transmit between people by both small and large droplets, whereas other nonrespiratory viruses, like EBOV, could theoretically be spread by large droplets because small droplets containing these viruses are not known to form in the human respiratory tract. It is unknown whether certain AGMPs generate either small or large droplets, or both. Therefore, depending on what aerosols are formed, AGMPs could potentially amplify a normal route of transmission for respiratory viruses or open up a new route of transmission for other viruses.
We can group possible AGMPs into two categories: procedures that mechanically create and disperse aerosols and procedures that induce the patient to produce aerosols (Figure 1 and Table 1). Procedures that irritate the airway, such as bronchoscopy or tracheal intubation, can cause a patient to cough forcefully, potentially emitting virus-laden aerosols, and both of these procedures are associated with the possibility of increasing the risk of SARS-CoV transmission among HCWs. The pressure on a patient’s chest during cardiopulmonary resuscitation can also induce a “cough-like force”, which was another possible source of SARS-CoV nosocomial transmission. Sputum is also routinely collected from patients for diagnostic purposes by cough induction, but it is not associated with nosocomial virus transmission.
In contrast to causing a patient to produce aerosols, AGMPs can also mechanically create and disperse respiratory aerosols through procedures such as ventilation, suctioning of the airway, or nebulizer treatment. Both manual ventilation, using a bag-valve-mask, and other forms of noninvasive ventilation (NIV), such as continuous positive airway pressure (CPAP), bilevel positive airway pressure (BiPAP), and high-frequency oscillatory ventilation (HFOV) are associated with SARS-CoV nosocomial transmission. Although the exact mechanisms of how these procedures create virus-laden aerosols in the respiratory tract remain unknown, it is possible that forcing or removing air from the respiratory tract could generate aerosols.
While AGMPs are traditionally thought of in regard to the generation of respiratory aerosols, AGMPs can also aerosolize infected fluids in other regions of the human body. Surgical techniques can aerosolize blood and possibly viruses. For example, infectious HIV-1 was found in the aerosols generated by surgical power tools, and a tracheotomy was associated with SARS-CoV transmission. Lasers can create plumes of debris that contain infectious aerosolized virus, as well. It is important to recognize the range of AGMPs and the circumstances under which they might be performed on infected patients. In order to associate certain AGMPs with nosocomial virus transmission, researchers need to test whether certain procedures generate aerosols with infectious virus, either through hospital sampling or laboratory procedures.
Among all infectious agents, those transmitted through aerosols are the most difficult to control. The speed of dispersion of airborne infectious agents makes them hard to contain and protect against, and the wide reach of susceptible hosts makes the control of airborne pathogens a priority for public and animal health officials.
Infectious agents travel associated with particles of various natures including fecal material, dust, debris, water, respiratory fluids and, specifically in buildings housing animals, with bedding and hair particles. The composition and size distribution of these particles will determine the location of deposition in the susceptible host, and influence the time the infectious agents can remain suspended in the air, the distance across which they can be transported, and the survivability and infectivity of the pathogens [2–4]. Thus, particle size is central to the epidemiology of airborne pathogens.
Particles of small size can remain suspended in the air for long periods, potentially exposing a large number of susceptible individuals, including those close to the source and those at greater distances. A spherical particle of 4 μm in diameter takes 33 min to settle 1 m in still air, compared to a 1μm particle that will take 8 h. Relative humidity, temperature and wind currents are the most important environmental factors that will determine the settling time of airborne particles that contain volatile components.
There is limited information in regards to the particle sizes with which infectious agents are associated. In humans, measures of particle size for influenza A virus (IAV) have been evaluated in controlled laboratory and health care settings, airplanes, daycares and households [6–10]. Distribution of IAV particle size varied between studies and settings, with IAV found in particles 1–4 μm in diameter (49% of particles) and particles > 4 μm (46% of particles) in health care facilities, and particles < 2.5 μm (64% of viral copies) detected in public places (i.e. health care, airplanes and day cares). Furthermore, dispersion models that have incorporated particle size information have indicated the plausibility of airborne transmission within these settings.
For pigs and cattle, the size distribution of particles associated with foot and mouth disease virus (FMDV) depended on whether aerosols were generated artificially or by experimentally infected animals. In the case of pigs affected by Aujeszky’s disease virus (ADV), particle size distribution varied based on days of infection. Specifically for IAV, no information was found for animals raised in agricultural environments, and the information is limited to animal models using ferrets or guinea pigs that assess the risk of IAV airborne transmission to people [14–16]. Overall, there is a lack of information on particle size distribution for pathogens affecting humans and animals, including zoonotic viruses. Furthermore, studies have been limited to viruses affecting the respiratory tract even though there is evidence that enteric viruses [i.e porcine epidemic diarrhea virus (PEDV), human-noroviruses, adenoviruses and enteroviruses] can also be transmitted through the air [17–20].
In this study we characterized the particle concentrations, size distributions and infectivity of three viruses that affect swine. These viruses were selected because of their differences in pathogenesis and modes of transmission. IAV, important because of its zoonotic potential, is shed only in respiratory secretions of pigs, and it is well established that it can be transmitted through aerosols. Porcine reproductive respiratory syndrome virus (PRRSV) is a systemic virus that can be shed in many body secretions, is exhaled in air, and has been detected in air as far as 9.1 km from swine herds. Lastly, PEDV, an emerging virus in North America, is an enteric virus that primarily replicates in large quantities in the small intestine, is largely concentrated in feces of diarrheic pigs and can replicate in alveolar macrophages. The information from this study may contribute to the understanding of airborne transmission of viruses of different pathogenesis and routes of transmission, which is necessary to fully prevent the spread of infectious diseases.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention.
Some epidemiological data support the possibility of both contact and airborne transmission of coronaviruses; however, the modes of transmission of MERS-CoV specifically are not completely understood. There is no consensus regarding precautionary recommendations for MERS-CoV; whereas the WHO recommends the contact and droplet precautions for all suspected cases and airborne precautions only for aerosol-generating procedures, the CDC advocates the use of airborne and contact precautions for all patient care activities. Several countries have taken the hard line of the CDC guidelines and recommend that direct healthcare workers should place all patients with a suspected infection in an isolated negative-pressure room and take adequate precautions by wearing two pairs of gloves, a disposable gown, and a face shield, and donning an N-95 respirator [28, 33, 34]. Currently, all suspected and confirmed cases should be treated in hospitals regardless of their medical condition, however transportation of patients to hospital requires additional emphasis, taking into account the small space of the vehicle and the close contact with the patient. The few guiding principles are that patients should be transported on a dedicated mission with the minimum number of crew members and should wear a surgical mask, if tolerated. The direct healthcare providers should wear the aforementioned PPE against contact and airborne transmission and avoid cough- or aerosol-generating procedures unless necessary. In addition, all transportations of patients should be coordinated with public health services.
Upon epidemiological investigation of 42 confirmed cases, 36 patients were directly admitted to AIIR, and 6 patients initially received care in non-AIIR facilities. Of the 413 HCWs caring for these patients before confirmation of SARS-CoV-2, 11 HCWs (2.7%) had been in close contact with unprotected exposure and required quarantine for 14 days. None of them was infected with SARS-CoV-2 by the end of the quarantine. Nosocomial transmission was not observed in these hospitalized patients.