June 22, 2020
Zeshan Qureshi1, Nicholas Jones2, Robert Temple3, Jessica PJ Larwood4, Trisha Greenhalgh,2 Lydia Bourouiba5
1 St Thomas’ Hospital, London, UK2 Nuffield Department of Primary Care Health Sciences, University of Oxford, Oxford, UK3 Somerville College, University of Oxford, UK4 St John’s College, University of Oxford, UK5 Massachusetts Institute of Technology, Cambridge, MA, USA
Correspondence to zeshan.qureshi
nhs.net or lbouro
Lay Summary by Mandy Payne, Health Watch
VERDICTThe 2-metre social distancing rule assumes that the dominant routes of transmission of SARS-CoV-2 are via respiratory large droplets falling on others or surfaces.A one-size-fits-all 2-metre social distancing rule is not consistent with the underlying science of exhalations và indoor air. Such rules are based on an over-simplistic picture of viral transfer, which assume a clear dichotomy between large droplets and small airborne droplets emitted in isolation without accounting for the exhaled air. The reality involves a continuum of droplet sizes & an important role of the exhaled air that carries them.Smaller airborne droplets laden with SARS-CoV-2 may spread up lớn 8 metres concentrated in exhaled air from infected individuals, even without background ventilation or airflow. Whilst there is limited direct evidence that live sầu SARS-CoV-2 is significantly spread via this route, there is no direct evidence that it is not spread this way.The risk of SARS-CoV-2 transmission falls as physical distance between people increases, so relaxing the distancing rules, particularly for indoor settings, might therefore risk an increase in infection rates. In some settings, even 2 metres may be too cchiến bại.Safe transmission mitigation measures depend on multiple factors related lớn both the individual và the environment, including viral load, duration of exposure, number of individuals, indoor versus outdoor settings, màn chơi of ventilation và whether face coverings are worn.Social distancing should be adapted và used alongside other strategies lớn reduce transmission, such as air hygiene, involving in part maximizing và adapting ventilation to lớn specific indoor spaces, effective hvà washing, regular surface cleaning, face coverings where appropriate và prompt isolation of affected individuals.
Bạn đang đọc: What is the evidence to support the 2
BACKGROUNDIt is well established that respiratory viruses can be transmitted via viral-laden mucosalivary respiratory droplets. These are expelled during exhalation, speech, và more forcefully during coughing & sneezing (categorised as ‘violent respiratory events’).1 Traditionally, transmission of respiratory disease was thought lớn occur via one of two distinct routes (và this classification is still being used): droplet route for large droplets and aerosol or airborne route for small droplets. The former assumes that large droplets fall on surfaces or on others, và contributes to surface contamination. The latter assumes inhalation of pathogen-bearing droplets invisible lớn the naked eye, & typically smaller than 5-10 microns in diameter. Based on this dichotomy framework, SARS-CoV-2 vi khuẩn is currently thought khổng lồ be spread by ‘tương tác and droplet’ transmission,2 though scientists are debating the possibility of airborne transmission.
Public health infection control measures are based largely on this classification, with different interventions recommended for large droplets & airborne droplets. Whilst some cấp độ of personal protective sầu equipment (PPE) và hvà hygiene are required in all cases, additional infection-control measures are recommended for airborne transmission in high-risk settings, such as the use of respirators & negative sầu pressure airborne individual isolation rooms.3 A parallel rapid Đánh Giá in this series, covering which clinical procedures are classed as aerosol generating, is due to lớn report shortly. Maintaining physical distance from other people may be effective at reducing transmission of both droplet và airborne spread of communicable disease. So-called ‘social distancing rules’ have sầu been implemented in many countries lớn reduce the risk of COVID-19 transmission, with early implementation of social distancing linked khổng lồ decreased incidence.4 5
Whilst this conceptual framework of droplet form size can be useful, the dichotomy between large droplets và small airborne particles emitted in isolation is an oversimplification. In fact, respiratory infections are transmitted via a continuum of droplet sizes embedded in a cloud of exhaled air containing those visible khổng lồ the naked eye (millimeters), lớn invisible ones, in the micron scale. This droplet size continuum & the cloud that carries them have sầu significant implications for mode of transmission. Social distancing rules are based on estimations of risk of droplet transmission in relation to isolated large droplet emission only. Therefore, if SARS-CoV-2 were only transmitted in large isolated droplets, this would imply shorter physical distancing measures would be sufficient khổng lồ reduce risk.
High rates of secondary infection have sầu been reported aước ao household members và cthua thảm contacts of people with COVID-19, who are likely khổng lồ get within 1-2 metres.6-9 The World Health Organization (WHO) have sầu suggested adopting a 1 metre social distancing policy, based primarily on the assumption that SARS-CoV-2 is transmitted in large isolated droplets.2 Individual countries have subsequently mix their own social distancing policies, with some countries (such as Spain và Canada) implementing a required 2-metre distance rule. The UK government’s recommendation is currently 2 metres but this is under Review at the time of writing.10 It is important that this distance is mix appropriately và flexibly. Too short a distance imposed too rigidly risks avoidable transmission, whereas too long is unnecessarily disruptive to society.
This reviews aims khổng lồ identify the evidence behind the 2-metre social distancing rule in the context of the still used large vs. small (droplet vs airborne) droplet kích thước dichotomy in route of transmission. It will look specifically at transmission risk in relation to lớn physical distance & air sampling studies around COVID-19 patients, but also wider evidence about whether airborne transmission should be considered a possible mechanism of SARS-CoV-2 spread.
What is the evidence for how far respiratory droplets of different kích cỡ travel?The origin of the 1-2 metre social distancing rule for communicable disease significantly predates COVID-19. A 1942 study by Jennison, referenced as early evidence for use of the 1-2 metre limit, used high-tốc độ exposure still photography to lớn detect atomized secretions & found that the majority of droplets were expelled within 1 metre.11 However, the giải pháp công nghệ was insufficient lớn capture smaller droplets & the chosen observation field for imaging was phối at 1-2 metres, meaning longer distance of droplet spread was not part of the study. Over time, limited epidemiongắn gọn xúc tích and simulated studies of selected infections such as rhinovirus và meningococcal disease have sầu generated some evidence khổng lồ tư vấn 1-2 metres of social distancing.12-14 However, these earlier studies also used methods that laông chồng accuracy compared khổng lồ current standards, particularly with respect khổng lồ air sampling. For example, Duguid et al used a method involving impact on glass slides và found no droplets less than 5µm were captured during coughing onto lớn exposed slides held within 6 inches of the face,15 but this result is not consistent with more recent data with modern air sampling methods of detecting droplet spread.
In 20đôi mươi, Bahl et al conducted a systematic review looking at reported horizontal distance travelled by respiratory droplets.11 This is useful as a proxy measure for how far associated viral particles might travel, & therefore risk of infection transmission in relation khổng lồ distance. Eight of 10 studies demonstrated a horizontal trajectory greater than 2 metres for particles up lớn 60µm. In one of the only studies using direct measurements involving human subjects in addition khổng lồ modelling, Bourouiba et al took & analysed direct high-tốc độ imaging of human sneezes và coughs. They noted the importance of the exhalation gas cloud in transporting all drops forward. They showed that although the largest droplets visible with the naked eye (order of millimeters) rapidly settled within 1-2 metres, the other droplets could be observed in air 6-8 metres away.16-18 This shows the potential for viral particles khổng lồ be projected extensively in a room within seconds of their emission.
Multiple studies have suggested possible spread beyond 2 metres from an index patient in the recent SARS, MERS và avian flu outbreaks.19 20 For example, Wong et al reported on transmission in relation lớn a SARS-CoV-1 outbreak amongst medical students exposed to a single patient in hospital.21 Of 27 students who entered the patient’s cubicle (used as a proxy marker for being within 1 metre), 10 developed the disease. However, 1/trăng tròn who denied entering the cubicle, & 4/18 who couldn’t rehotline if they had entered the cubicle, also developed the disease, which could suggest either secondary transmission or possible transmission over longer distances.
What factors influence the distance respiratory droplets spread?Given droplet sizes are on a continuum rather than binary large or small, the distance droplets travel will also be over a continuous range & influenced by a number of factors, other than droplet kích cỡ alone. For example, recent research by Bourouiba et al has shown by direct quantification, modelling, and validation on human subjects that ‘violent respiratory events’, e.g. coughing and sneezing, generate a warm, moist and turbulent gas cloud with forward momentum. This can significantly extkết thúc the distance travelled by a virus in a room within seconds & independent of background ventilation or airflow. This phenomenon and other developments in our understanding of exhalation dynamics are not accounted for in the earlier modelling studies of droplet transmissions on which the current dichotomous classification of large versus small droplets is based.16
Even volume of speech may impact on droplet spread & subsequent risk of transmission, making the process of predicting mode of transmission problematic.22 Clusters of coronavirus have sầu occurred during prolonged ‘violent exhalation events’ such as singing or fitness dance classes in confined locations.23 24 For example, Hamner et al report that a two and a half hour choir rehearsal with one symptomatic person led khổng lồ 32 confirmed and 20 probable COVID-19 cases ahy vọng the 61 singers, even though all singers avoided any direct physical tương tác.24
Viral shedding (higher with coughing/sneezing) & factors related lớn airflow in indoor environments, such as ventilation, may increase droplet spread. Nishiura et al used tương tác tracing khổng lồ collect secondary transmission data from 110 index cases with COVID-19 across 11 clusters in nhật bản. All clusters were linked to lớn indoor spaces, including fitness gyms and a restaurant boat. The authors report that the odds of transmission in an enclosed environment were 18.7-fold higher than in an outdoor environment.25 Other indoor case clusters have sầu been reported in gyms, churches, hospitals and elderly care settings.26 27 Conversely, facemasks may help limit airborne droplet transmission.28 Such case reports require further investigation but imply environmental factors are important in addition to physical distance in determining risk of transmission.
This indirect evidence illustrates that safe social distancing limits differ widely between settings, with outdoor environments likely associated with lower risk of transmission at a given distance. Staggered social distancing rules alongside other public health interventions may be required to lớn recognise the importance of the environmental context in determining transmission risk.
What is the evidence that there is live vi khuẩn in these different sized droplets at different distances?Respiratory virut RNA is detectable in both large droplets that settle on surfaces và airborne droplets after breathing or ‘violent respiratory events’.29 Both animal and human volunteer influenza inoculation studies have sầu suggested that deep inhalation (which occurs khổng lồ a far greater extent with aerosols than droplets) may result in similar or even greater infection rates compared to intranasal large droplet inoculation.30-32 However, the relative contribution of the airborne route to actual transmission remains a subject of debate.17 33 34 This may in part be due to variation in host, viral, and environmental factors for each interaction.35 Factors suggested include vi khuẩn concentration in respiratory fluid, levels of pollution or particulate matter in the air, humidity, temperature, indoor versus outdoor environments, symptomatic versus asymptomatic hosts, and an individual’s baseline susceptibility khổng lồ infection.
SARS-CoV-2 is present in sputum.36 Van Doremalen et al, analysed SARS-CoV-2 across 10 experimental conditions in five sầu environments and showed that the vi khuẩn is also stable in air for at least 3 hours,9 with others suggesting it may be stable for up to lớn 16 hours.37 Reports from cluster outbreaks such as the choir practice provide indirect evidence there is live SARS-CoV-2 in respiratory droplets.24 There is also additional indirect evidence suggesting potential airborne transmission. One study showed that SARS-CoV-2 is deposited deep in the airways of hospitalized patients.36 This tends to be associated with an airborne route of transmission, whereas diseases of the upper respiratory tract tend to be associated with rapidly settling droplets and surface contamination. Additionally, asymptomatic spread of coronavirut has been confirmed in several studies,38-40 which is consistent with airborne transmission as larger visible droplets are disproportionately emitted in coughing & sneezing.41 42
What is the evidence that 2 metres is an adequate distance khổng lồ reduce SARS-CoV-2 transmission?
Search strategyTo determine what evidence there is lớn tư vấn the 2-metre social distancing rule specific lớn SARS-CoV-2, we conducted a tìm kiếm of PubMed, MedRxiv, LitCOVID và Google scholar, using the terms listed in the appendix. We included studies reporting transmission risk in relation lớn distance across any setting. We also included studies reporting airborne sampling for SARS-CoV-2, as we felt these might provide information as to the potential spread of the vi khuẩn in relation to lớn distance. The search was run from inception to 17th June 2020. From 3,549 papers identified in the tìm kiếm & an additional 58 studies through forward and backward citation checking, we reviewed 120 full texts. From these we included 25 studies directly reporting SARS-CoV-2 transmission risk in relation khổng lồ physical distance. We also included additional texts in the Review that report distances in droplet spread not specific khổng lồ SARS-CoV-2, though these were not the focus of the search.
Overview of studiesAside from one systematic Reviews across settings, we categorise results inkhổng lồ community (n=10) and hospital-based studies (14), given disease prevalence and severity are likely to differ significantly. Community settings included cruise ships (2), household contacts (3), a restaurant (1), a shopping mall (1) a medical conference (1) a multi-storey mixed commercial & residential-use building (1), và a multi-site study (1). In addition to lớn the systematic nhận xét (1), individual studies were based in China (10), the USA (4), Singapore (2), Germany (2), the United Kingdom (1), South Korea (1), Taiwan (1), xứ sở của những nụ cười thân thiện (1) & two aboard cruise ships. None were specifically mix outdoors or in schools. Additionally, although the demographics of several of the studies are unclear, none look specifically at children or babies.
Quality assessmentMost of the included studies had not yet undergone peer nhận xét và many were felt to lớn be at risk of bias due to lớn small numbers of study participants, và methods lacking transparency or reproducibility. There was heterogeneity across the studies in terms of methods, population & retìm kiếm question. Many studies were of retrospective thiết kế, & hence at risk of regọi bias in terms of distance lớn infected individual and selection bias in terms of identifying recent contacts, particularly those that relied on contact tracing alone. Confounding variables, such as disease severity, time since symptom onphối and contact with other individuals, were rarely reported. Publication bias was not formally assessed but must be considered, particularly in the reporting of clusters of cases. All of these factors may be important in understanding the variation in reported results.
In the air particle studies, only two incorporated the capacity to directly measure the infectivity of coronavirus, rather than just the presence of viral RNA in the air. Heterogeneity specific khổng lồ these studies included differences in hospital cleaning and ventilation, variability in air sample volume & sample handling for viability demo. These studies also tended khổng lồ include insufficient discussion of systematic calibrations for sampling instrumentation or sensitivity of viability lớn sample collection and handling methodologies. This makes it difficult khổng lồ interpret the results objectively và compare results between sampling studies.
A recent systematic đánh giá & meta-analysis published in The Lancet assessed the evidence for reducing risk of SARS-CoV-2 transmission.43 They aimed lớn investigate optimum physical distancing for avoiding person-to-person coronavirut transmission, & to lớn assess the impact of face masks & eye protection on preventing transmission. Studies of any kiến thiết in any setting were included if they reported on these outcomes aao ước WHO-defined confirmed or probable COVID-19, SARS, or MERS, published before 3rd May 2020.
However, this reviews draws largely on evidence from SARS và MERS with only seven studies included of COVID-19, five of which were non-peer reviewed pre-prints & one a correspondence piece. There was significant heterogeneity between the included studies in terms of setting, indoor & air conditions, degree of physical distancing & identification of cases, which makes it difficult lớn draw conclusions as to the safety of respective sầu distances. As with the other observational studies we include in our reviews, there is risk of rehotline bias in terms of people remembering how cchiến bại they were khổng lồ contacts và selection bias in terms of regular contacts of infected cases being more likely to lớn be included. Such limitations mean there is low or very low certainty in these summary findings và thus no methodological ability lớn distinguish between routes of transmission in an indoor space.
SARS-CoV-2 transmission risk across all settingsIn their Lancet systematic Reviews, Chu et al reported that closer liên hệ distance is associated with an increased transmission risk of SARS-CoV-2, across study settings. In a sub-group analysis focused on SARS-CoV-2, the relative sầu risk of developing COVID-19 aao ước people who had been in “close” compared khổng lồ “distant” contact with an infected case was 0.15 (95%CI 0.03 to lớn 0.73). However, the threshold for ‘further distance’ varied from anything over direct contact in some studies khổng lồ 2 metres in others, meaning a cthua tương tác in one could be a distant liên hệ in another. Chu et al estimated this key physical distance for some studies where it had not been explicitly reported. Furthermore, there was no real accounting for other variables affecting transmission risk beyond just social distancing, and that may explain some of the variability between studies. However, in a meta-regression of change in relative sầu risk of developing SARS-CoV-1, SARS-CoV-2 or MERS in relating to lớn increasing distance, the risk of being infected is estimated lớn be 13% for those within 1 metre, but only 3% beyond that distance. The authors go on to lớn conclude that there is good evidence khổng lồ support physical distancing of at least 1 metre, but 2 metres may be more effective sầu, whilst acknowledging a range of factors influence transmission risk. This analysis assumes transmission risk in relation to distance is fixed và does take tài khoản of important variables such as duration of exposure or the indoor air and environment. The meta-analysis also found some evidence khổng lồ tư vấn facemasks (aOR 0·15, 95%CI 0·07 to lớn 0·34 with stronger associations with N95 or similar respirators) & eye protection (aOR 0·22, 95% CI 0·12 lớn 0·39) for reducing transmission of coronavirus. Eye protection benefits would be consistent with the airborne route of transmission, given that smaller particles are likely to lớn remain present in the air và be absorbed via the conjunctival surface.44
SARS-CoV-2 transmission risk in community studiesThe ten community studies in our sample included five sầu that retrospectively analysed the impact of physical distance on COVID-19 outbreaks, three liên hệ tracing studies và two studies of air sampling around COVID-19. Five studies reported clusters of cases aước ao individuals who had prolonged exposure to an infected individual within a confined space,38 45-48 và a single modelling study also suggests this may be important for transmission.49 Spouses and cthất bại household contacts also appeared khổng lồ be at increased risk, compared to community contacts,50 51 though interestingly air sampling studies in the houses of people with COVID-19 were negative.52 Although some studies suggested possible transmissions at distances in excess of 2 metres,47 53 there wasn’t sufficient description lớn rule out transmission via cthua kém or direct liên hệ in any of the studies. We describe these studies in more detail below.
Retrospective analyses of COVID-19 outbreaks
Li et al reviewed study analysed an outbreak of 10 new cases of COVID-19, all of whom were infected in a single sitting at a restaurant in Guangzhou, China.47 The 10 cases across three families all sat in adjoining tables at one end of the restaurant. Transmission is believed to lớn have taken place between these individuals despite no significant cthất bại contact between the involved families on video clip analysis. Distances between index và patrons infected were noted khổng lồ be up khổng lồ 4.6 metres. The pattern of transmission was compatible with flow patterns of indoor ventilation in the restaurant, which they were sitting beneath. Transmission via touch cannot be excluded but there was a specific pattern of infection along the downstream flow line from the ventilation source and no cases were reported from other tables within the restaurant. Both the pattern of transmission, & the distance of separation tư vấn airborne transmission.
Hijnen et al noted the likely spread of COVID-19 at a dermatology meeting of 14 people, of which 12 subsequently tested positive sầu including the index case.38 At the meeting, individuals sat 2.6 metres away from the index individual over a two-day meeting. They also shook hands và shared taxis, giving opportunity for closer contact.
Cai et al analysed a 17-person outbreak of COVID-19 at a shopping mall in China.53 They found that several individuals who contracted the disease worked on different floors to the index case, raising the possibility of longer distance transmission through air. This might also be explained by shared routes to lớn work, elevators, spread from asymptomatic individuals, congregation of staff, or contaminated customers moving between floors.
Park et al investigated an outbreak of COVID-19 cases at a single 19 storey building, containing a phối of residential & commercial units.45 Liên hệ tracing was initiated after the first index case was detected, leading to lớn 97 confirmed cases of 1,143 tested, 89 (91.7%) of whom were symptomatic. Most of these positive cases (n=94, (96.9%)) worked in an 11th floor Hotline centre, which had 216 employees in total. Although the index case did not visit this điện thoại tư vấn centre or floor, the second case-patient was an employee at the Call centre. There was a secondary attack rate of 34 (16.2%) in the households of the positive sầu cases.
Xu et al published an analysis (not yet peer reviewed) of the COVD-19 outbreak aboard the Diamond Princess Cruise Ship, where 696 people were infected of the 3711 aboard.46 They included 197 initial symptomatic cases as well as an additional 146 subsequent passenger cases, 129 of whom had been in ‘cđại bại contact’ with one of the original infected individuals, defined as staying within the same state room, and 17 who developed COVID-19 but had not been in close contact with an index case.46 Their analysis suggests these 17 people were infected before quarantine measures were introduced, suggesting there was no distant spread following quarantine. This was a retrospective Mã Sản Phẩm, và provides limited evidence as lớn safe distancing limits and routes of transmission in indoor spaces.
Air sampling studies
The second cruise ship study, also non-peer reviewed, aimed to lớn establish whether contaminated environmental surfaces were important in determining risk of SARS-CoV-2 transmission.54 The study team directly sampled air in cabins of those with and without COVID-19, finding no evidence of coronavirut across 14 air samples, but evidence of it on surfaces, such as bed pillows & around the nhà vệ sinh, based on positive sầu RNA results. However, samples were taken up to lớn 17 days after the residents had left their cabins & the negative results could be a reflection of this long lag time before testing. Döhla et al analysed air samples from 21 households under lockdown where a family thành viên had tested positive sầu for COVID-19 and reported all samples were negative sầu.52 Timing in relation lớn cleaning and distance from the infected person was unclear.
Liên hệ tracing studies
Doung-ngern et al conducted a retrospective sầu case-control study in Vương Quốc Của Những Nụ cười using tương tác tracing to lớn identify 1,050 asymptomatic individuals who had been in cthua tương tác with 18 primary index patients with COVID-19.48 Close contacts included household members or non-household contacts who had been within 1 metre of the index patient for longer than 5 minutes. Most people contacted the index cases at a boxing match (n=645), night club (n=374) or at a state enterprise office (n=31). Of these initially asymptomatic people, 211 (20%) were later diagnosed with COVID-19. People who did not come within a 1 metre distance of the index case were at significantly lower risk of developing COVID-19 (adjusted odds ratio 0.15, 95%CI 0.04 to 0.63), whilst facemasks and handwashing were also associated with reduced risk.
In a prospective sầu study among mỏi a Taiwanese population, Cheng et al included 32 confirmed cases of COVID-19 and their 1,043 recent contacts to lớn determine the secondary attack rate (i.e., spread of disease onto lớn other cthua kém contacts) & determinants of transmission. There were 15 cases identified as secondary attacks in close contacts, defined as people who had spent 15 minutes or more in face-to-face tương tác with the infected individual. There was no statistically significant difference in secondary attaông chồng rates between household contacts (13.9%, 95%CI 4.7 to 29.5%) and family contacts not living in the same household (8.5%, 95%CI 2.4 to lớn đôi mươi.3%).51 There were no secondary cases reported amongst more distant contacts, including healthcare workers.
Burke et al identified nine cases with COVID-19 early in the outbreak in the USA, & 404 of their cthua trận contacts who agreed to participate in monitoring. Of these cthua contacts, 159 had 1 or more respiratory swab taken. There were only 2 cases of secondary transmission, both among muốn spouses, giving a secondary attaông xã rate of 13% (95%CI 4 lớn 38%) among muốn the 15 household contacts. Exposure khổng lồ the infected individuals was longer for these two spouses than the 13 individuals who did not develop COVID-19.50 None of the healthcare professional or community contacts developed swab-positive sầu COVID-19, though many had suspected symptoms, potentially calling inlớn question the accuracy of the tests used.
SARS-CoV-2 transmission risk in hospital settingsOf the fourteen hospital studies, nine analysed air samples around patients with confirmed COVID-19 as a proxy measure for possible airborne spread (Table 1). Seven of these studies reported positive sầu airborne samples for SARS-CoV-2, including two at distances of 2 metres or greater from the source patient. Although the presence of SARS-CoV-2 in airborne particles does not confirm transmission at distance, it does demonstrate the extent to lớn which viral-laden droplets may travel in air from an infected individual, consistent with what is known about high momentum cloud exhalation range55 and not consistent with the ballistic large droplet route, in which the all contaminated droplets would fall on surfaces within the 1-2 m vicinity of the patient.
Air sampling studies
Zhou et al investigated airborne spread of SARS-CoV-2 at a multi-site London Hospital.56 Viral RNA was detected on 14/31 (38.7%) air samples và 114/218 (52.3%) of surface samples, with more frequent positive results in areas occupied by patients with COVID-19. No vi khuẩn was deemed viable (i.e. capable of being replicated) from any site. However, this may have sầu been due khổng lồ the time lag between droplets deposition and culture, and small air sample volumes. Liu et al found positive sầu SARS-CoV-2 air samples across 60% of tested sites in two dedicated COVID-19 hospitals.57 Chia et al conducted air sampling in the rooms of three patients with confirmed COVID-19. Positive results were obtained from the two patients who were on day 5 of symptoms but not the patient who was on day 9.58 Ma et al analysed air and surface samples of local indoor environments of COVID-19 recruited subjects in addition to their exhaled breath condensate. They found positive sầu surface (5.4% out of n = 242) and air samples (3.8% out of n = 26) in addition khổng lồ 103-105 viral RNA copies/min in exhaled breath associated with a positive sầu sampling rate of 16.7% out of n = 30 samples.59 Santarpia et al reported positive sầu samples collected from air samplers worn by sampling personnel in the vicinity of patients và at fixed distance locations from the patient, even when the patient did not cough. Additionally, they reported positive air samples from fixed collectors located at least about 2 meters away (6 feet).60 They also assessed viral viability, reporting signs of viability via viral propagation và using various indicators of viral replication. Indication of replication competence was reported for windowsill và hallway samples despite small volume sample preventing a full replication study. They noted that the small volume of recovery was problematic.60 Indeed, although there were no viable virus samples in most other air samples across the studies, there were also no viable vi khuẩn samples on surface samples either. Hence, laông xã of viability in these studies cannot be used to discriminate between routes of transmission at this stage.
In contrast, Ding et al analysed 46 air samples from a hospital in Nanjing, including from isolation rooms of people with COVID-19 và found only one weakly positive result, which came from a ward corridor. However, they found that the exhaled condensate and two expired air samples from patients were also negative sầu.61 It is also unclear whether the isolation rooms were occupied at the time of sampling, which might explain these negative sầu environmental sampling results. Ong et al. also reported negative sầu air samples taken in the vicinity of three patients with COVID-19, but positive samples in the room air exhaust outlets, consistent with the airborne route & suspension và clearance of airborne virus-laden particles by ventilation.62 Viral particles in air-vents is not consistent with the hypothesis that the virus would only be contained in large drops deposited ballistically on surfaces in the 1-2 m patient vicinity. Wu et al also detected no positive air samples ahy vọng 44 collected on a medical ward caring for people with COVID-19.63 However they discussed the limitation of their findings with the need for larger air sampling volume.
Only two sampling studies explicitly commented on distance. Guo et al reported airborne samples positive for SARS-CoV-2 up to lớn 4 metres from the patient,64 và Sanatarpia up lớn at least 2 metres.60
Wong et al in Hong Kong undertook liên hệ tracing around a patient with COVID-19, receiving oxygen and treated for pneumonia.65 Although 52 individuals came into lớn tương tác with the index case & subsequently developed fever or respiratory symptoms, all tested negative for SARS-CoV-2. No explanation is given for the high rate of symptoms in these contacts và the results must be considered with caution.
Heinzerling et al undertook a study in California, reporting that 3/33 healthcare professionals who came within six feet of an index patient with COVID-19 later developed the infection và tested positive sầu.66 Of these three, two had frequent direct tương tác with the index patient, including during aerosol generating procedures, but wore no facemask, eye protection or gown. The third staff member wore a facemask and gloves most of the time, but did not wear eye protection.
Bai et al reported 12/42 healthcare professionals in a hospital in Wuhan who had contact with either the index patient with COVID-19 or an affected colleague (distance undefined) developed the infection, compared khổng lồ 0/76 colleagues without such tương tác.67 It is unclear khổng lồ what extent each group complied with PPE advice. Burke et al investigated 126 healthcare professionals who had tương tác with nine patients with COVID-19. None developed COVID-19, even though 76 provided direct patient care, of which only 43% reported using appropriate PPE.50 Cheng et al studied 301 healthcare professionals exposed to lớn 32 confirmed COVID-19 patients, defined as being within 2 metres without appropriate PPE, yet none tested positive sầu for COVID-19.
Table 1. Summary of evidence from hospital-based studies looking at airborne transmission of SARS-CoV-2
|Study||Year||Setting||COVID-19 Patients||Air Samples||Air Sample Volume||Method of detection||Peer Review||Viability Test (can the vi khuẩn replicate)||SARS-CoV-2 Positive Air Samples||Possible evidence for airborne transmission||Comments|
|Liu57||2020||Wuhan, China||Unknown, though the two hospitals were exclusively used for people with COVID-19 during outbreak||30 sites across two hospitals including public sites||Total sampling air volumes: 1.5 m³ lớn 8.9 m³, 5L/min rate.|
Sampled between 5 hours & 7 days
|PCR||Yes||No||21 positive samples across 35 sites (60%), though 4 sites were positive sầu at first round of sampling and negative sầu at second||Yes||Distance of air samples from patients unclear|
|Guo64||2020||Wuhan, China||39||40 ICU samples & 16 general ward samples||300l/min (30 minutes)||PCR||No||No||In ICU, 5/14 samples near the patient, 8/18 approximately 2.5 metres from the patient, and 1/8 samples approximately 4 metres from the patient. In general ward 2/11 positive near the patient and 0/5 at 2.5 metres distance.||Yes||Unclear why higher percentage of positive sầu results at 2.5 metres from the patients compared to lớn immediately by the patients.|
|Santarpia60||2020||Nebraska Medical Centre, USA||11 rooms with 13 COVID-19 patients và hallways||31||50l/min (15 minutes)||PCR||No||Yes||63.2% (including 2/3 where the air sampling was at least six feet (about 2 metres) from the patient & 58.3% in hallways. All non-viable.||Yes||Distance of air sample from the patient was not uniformly recorded for any samples other than the three reported. Indication of viral replication competence in samples identified, with limitation in volume sampled discussed as problematic.|
|Ding61||2020||Nanjing, China||10||46||Multiple groups:|
b) 50l/min (20 mins)
c) 500l/m (2 min)
d) 500l/m (20 min)
e) 14l/min (30 min)
|PCR||No||No||2% i.e. 1 sample, which was weakly positive||Yes||The significance of weakly positive is unclear. Distance of air samples from patients unclear.|
|Chia58||2020||Singapore||3 rooms||18||3.5l/min (4 hours)||PCR||No||No||Positive samples in 2 of 3 rooms||Yes||Distance of air samples from patients unclear|
|Ong62||2020||Singapore||3||Unclear||5/l min (4 hours)||PCR||Yes||No||All airborne samples were negative sầu. However, positive swabs were collected from 13/15 room sites, including 2/3 at the air outlet bạn.||Yes given positive samples at air outlet fans, although all air samples were negative||Inconsistent methodology between samples reported. Majority of samples collected immediately after room cleaning|
Distance of air samples from patient unclear
|Zhou56||2020||London, UK||7 Clinical areas và 1 public area of the hospital||31||100L/min (10 mins)||PCR||No||Yes||38.7% suspected positive sầu (at least one of two samples positive), 6.4% positive (both samples positive). All 8 areas produced at least one positive. All non-viable||Yes||Short sampling time.|
Distances from patients unclear.
Positive sầu samples more comtháng in areas where COVID-19 patients were being cared for
|Wu63||2020||Wuhan No.7 Hospital, China||17 hospital areas including intensive sầu care||44||As per “Hygiene Standards for Disinfection in Hospital” in China||PCR||Yes||No||0%||No||Small air sample volumes acknowledged as a problem|
Distance of air sample from patients unclear
|Ma59||2020||Beijing, China. (Two hospitals và multiple quarantine hotel rooms)||Samples taken from areas associated with 35 recruited COVID-19 patients. Hospital and quarantine khách sạn environments including corridors, khách sạn rooms, ICU emergency room and CT room & clinical observation room samples.||26||15L/min in enclosed environments|
400L/min in corridors
|PCR||No||No||3.8%||Yes||Also sampled exhaled breath of 30 patients, 16.7% positive sầu. Distance of air samples from patient unclear. Rooms well ventilated (open windows or negative pressure systems)|
CONCLUSIONSThe longstanding dichotomy of large droplet versus small airborne droplet transmission is outdated và SARS-CoV-2 may be present and stable in a range of droplet sizes, which will travel across a range of distances, including some beyond 2 metres.The majority of existing evidence specific to SARS-CoV-2 is observational và non-peer-reviewed, with significant heterogeneity in terms of populations, study settings, sample collection methods & primary outcome. Determining the relative risk of SARS-CoV-2 at different distances is therefore difficult from such studies.Evidence from community studies suggest prolonged exposure in an enclosed space, with unknown information about distancing, may be linked khổng lồ clusters of cases, particularly in the context of activities such as choirs, sports events or fitness gyms.Increasing physical distance is associated with decreasing risk, so easing restrictions from 2 lớn 1 metre may result in a significant increase in risk if other measures are not taken.Other factors such as duration of time spent with others in an indoor space, e.g. at work in a confined office và the indoor air conditions are as important khổng lồ account for in the estimation & mitigation of risk.Single thresholds for social distancing, such as the current 2-metre rule, over-simplify what is a complex transmission risk that is multifactorial. Social distancing is not a magic bullet lớn eliminate risk. A graded approach to physical distancing that reflects the individual setting, the indoor space và air condition, and other protective factors may be the best approach lớn reduce risk.Other important factors lớn take account when considering safe social distancing (which were beyond the scope of this reviews khổng lồ cover in depth) include host viral load, duration of exposure, number of infected individuals, indoor versus outdoor settings, air ventilation, wearing of PPE including facemasks, effectiveness and type of cleaning measures, individual susceptibility lớn infection, và activities that project airborne particles over greater distances in exhaled gas clouds, such as singing, coughing or heavy breathing.Social distancing should therefore be used in combination with other strategies khổng lồ reduce transmission risk, including h& washing, regular surface cleaning, PPE and face coverings where appropriate, strategies of air hygiene, and isolation of affected individuals.
Disclaimer: The article has not been peer-reviewed; it should not replace individual clinical judgement and the sources cited should be checked. The views expressed in this commentary represent the views of the authors and not necessarily those of the host institution, the NHS, the NIHR, the Department of Health and Social Care or the Centre for Evidence-based Medicine. The views are not a substitute for professional medical advice.
This article was nhận xét và amended June 24th 2020 with a change in the conclusion supporting further distance as opposed lớn a precise distance measure
AUTHORSZeshan Qureshi is a doctor based at St Thomas’ Hospital, and a previous Academic Clinical Fellow in Global Health.Niông xã Jones is a GPhường. and Welcome Trust Doctoral Research Fellow at the Nuffield Department of Primary Care Health Sciences, University of Oxford, UKRobert Temple is a Medical Student at Somerville College, University of Oxford, UKJessica PJ Larwood is a Medical Student at St John’s College, University of Oxford, UKTrisha Greenhalgh is Professor of Primary Care Health Sciences at the Nuffield Department of Primary Care Health Sciences, University of Oxford, UKLydia Bourouiba is Associated Professor at the Massachusetts Institute of Technology (MIT) và director of The Fluid Dynamics of Disease Transmission Laboratory, MIT, Cambridge, MA, USA.
We used the following tìm kiếm terms:
|(coronavirus OR covid-19) AND (transmission OR transmissability) AND (“social distancing” OR “physical distancing”) AND (1m OR 2m OR 3m OR 4m OR 5m OR 10 m OR metre OR metres OR meter OR meters)||Google & GoogleScholar|
|(coronavi khuẩn OR covid-19) AND (“two metre rule” OR “two meter rule” OR “2 metre rule” OR “2 meter rule” OR “six feet rule” OR “six foot rule” OR “6 feet rule” OR “6 foot rule”)||Google & GoogleScholar|
|((metre* OR meter* OR foot OR feet OR “1m” OR “2m” OR “3m” OR “4m” OR “5m” OR “3ft” OR “6ft” OR “9ft” OR “12ft” OR “15ft”) AND ((coronavirus*||PubMed|
|(social distanc* OR physical distanc* OR workplace distanc* OR school distanc*) AND (transmission OR transmissability OR spread)||LitCOVID|
|(social distanc* OR physical distanc* OR workplace distanc* OR school distanc*) AND (airborne OR droplet*)||LitCOVID|
|(social distanc* OR physical distanc* OR workplace distanc* OR school distanc*) AND (meter* OR metre* OR feet OR foot)||LitCOVID|
|“”(coronavi khuẩn OR covid-19) AND (distance OR distancing) AND (metre OR meter OR foot OR feet)”||medRxiv|
|“social distancing” (match phrase words) and abstract or title “coronavirut OR covid-19” (match any words)||medRxiv|
|title “rapid review” (match phrase words) and abstract or title “social distancing”||medRxiv|
|((aerosol* OR droplet*) AND (distance OR meter* OR metre* OR foot OR feet OR spread* OR dispers* OR transmi*)) AND ((coronavirus*||PubMed|