key: cord-0768981-75fgaj71
authors: Byrne, A. W.; McEvoy, D.; Collins, A.; Hunt, K.; Casey, M.; Barber, A.; Butler, F.; Griffin, J.; Lane, E.; McAloon, C.; O'Brien, K.; Wall, P.; Walsh, K.; More, S.
title: Inferred duration of infectious period of SARS-CoV-2: rapid scoping review and analysis of available evidence for asymptomatic and symptomatic COVID-19 cases
date: 2020-04-30
journal: nan
DOI: 10.1101/2020.04.25.20079889
sha: 592458f7eadad2c7a49452e6868c2ab6c3666d3f
doc_id: 768981
cord_uid: 75fgaj71

Objectives: Our objective was to review the literature on the inferred duration of the infectious period of COVID-19, caused by SARS-COV-2 virus, and provide an overview of the variation depending on the methodological approach. Design: Rapid scoping review. Literature review with fixed search terms, up to 1st April 2020. Central tendency and variation of the parameter estimates for infectious period in (a) asymptomatic (b) symptomatic cases from (i) virological studies (repeated testing), (ii) tracing studies (iii) modelling studies were gathered. Narrative review of viral dynamics. Information sources: Search strategies developed and the following searched: PubMed, Google Scholar, MedRxiv, BioRxiv. Additionally, the Health Information Quality Authority (Ireland) viral load synthesis was utilised, which screened literature from PubMed, Embase, ScienceDirect, NHS evidence, Cochrane, medRxiv and bioRxiv, HRB open databases. Results: There was substantial variation in the estimates, and how infectious period was inferred. One study provided approximate median infectious period for asymptomatic cases of 6.5-9.5 days. Median pre-symptomatic infectious period across studies varied over <1-4 days. Estimated mean time from symptom onset to two negative RT-PCR tests was 13.4 days (95%CI: 10.9-15.8), but was shorter when studies included children or less severe cases. Estimated mean duration from symptom onset to hospital discharge or death (potential maximal infectious period) was 18.1 days (95%CI: 15.1-21.0); time to discharge was on average 4 days shorter than time-to-death. Viral dynamic data and model infectious parameters were often shorter than repeated diagnostic data. Conclusions: There are limitations of inferring infectiousness from repeated diagnosis, viral loads, and viral replication data alone, and also potential patient recall bias relevant to estimating exposure and symptom onset times. Despite this, available data provides a preliminary evidence base to inform models of central tendency for key parameters, and variation for exploring parameter space and sensitivity analysis. Some current models may be underestimating infectious period.

Results 191

Overall, 65 parameter estimates were harvested from 48 papers (Tables 1, 2, 3) . 193

Infectious period for asymptomatic cases (T2) 194

The overall distributions and point estimates from studies for T2 are presented in Figure 1 and Table  195 1. 196

Two virological studies reported on infectious period based on serial diagnostic testing, for 197 asymptomatic cases, were found to have informative data. One of these studies reported on only 198 one asymptomatic case, with exposure to negative tests being 11 days (Zhou et al, 2020 ). This 199 duration should be considered an over-estimate, given that a latent period is not taken into 200 consideration. Hu et al. [7] tracked infections of close contacts to infected persons and considered 201 patients asymptomatic at time of diagnosis. Infectious period was defined as time from diagnosis to 202 the first of two clear tests, providing a median duration of 9.5 days (n=24) range: 1 -21; 3.5-13.0 203

Importantly, Hu et al. [7] found that the infectious period was different between those who 205 subsequently exhibited symptoms (i.e. pre-symptomatic) and those who did not: The median 206 duration for asymptomatic infectious was 6.0 days (IQR: 2.0 -12.0; N=19). This was reduced to 4.0 207 days (2.0 -15.0) for cases that were asymptomatic without abnormal computed tomography (CT) 208 scans (n=7). 209

Two tracing studies provide informative data (Table 1; Modelling studies that have attempted to fit differing parameters depending on the severity of 218 symptoms have used differing nomenclature, for example asymptomatic, "mild" or subclinical cases 219 (Table 1) . [14, 15, 26, 27] Two papers by Davies and colleagues [14, 15] model this parameter as a 220 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . gamma distribution with a mean periods of 5-7 days (Fig. 2) ; importantly, these papers assume 221 infectious period is the same for asymptomatic and symptomatic cases. 222

Pan et al. [3] and Hoehl et al. [28] describe the cases of two individuals tracked and serially tested by 224 real-time reverse transcriptase polymerase chain reaction (RT-PCR) after being exposed to a patient 225 with confirmed infection. In the latter study, the virus was isolated from samples, indicating 226 transmission potential. 227

Four studies from China, Germany and Singapore provide informative data through tracing infections 228 from cluster of infections, and through infector-infectee pairs ( Table 2) . [4, 9, 29 ,30] These papers There was high heterogeneity across studies (Cochrane's Q; p<0.001; I 2 >75%). A random effects (RE) 250 meta-regression model suggested significant variation depending on whether studies included 251 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. [29] estimated that the proportion of all transmission that was pre-symptomatic was 44% (95% CI, 283 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. there was no right censoring (that is, transmission being halted through isolation or hospitalisation; 299 gamma distributions of mean 5 or 7 days). Tuite et al. [26, 39] also assumed the same duration for 300 "mild" and "severe" symptomatic cases (6-6.5 days). 301 302 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. Hence, lower Ct values infer higher viral loads. The authors report on a patient without symptoms, 320 but with positive nasal swabs (Ct values, 22 to 28) and throat swabs (Ct values, 30 to 32) testing 321 positive on days 7, 10, and 11 after contact. Importantly, the authors suggest "the viral load that was 322 detected in the asymptomatic patient was similar to that in the symptomatic patients." 323 Furthermore, Kimbell et al. [62] report that Ct values between asymptomatic (21.9 to 31.0), pre-324 symptomatic (15.3 to 37.9), and symptomatic cases (18.6 to 29.2) within a nursing home 325 environment did not differ significantly. To et al.

[59] present data on temporal profile of viral load 326 from saliva samples, and found that median initial and peak viral loads in severe cases were non-327 significantly higher (p>0.5) by approximately 1 log10 higher than those in mild cases. Liu This lack of pre-symptomatic data may result in left truncation of the risk distribution associated 330 with viral load and shedding. Therefore, the typical timing of peak viral shedding (whether prior to, 331 at, or after onset), and it's impact on transmission, is still uncertain. He et al. [29] reported highest 332 viral load at symptom onset from patients sampled in a hospital in China. Furthermore, the author's 333 estimate using a separate infector-infectee dataset (n=77) that 44% (95% CI: 25-69%) of infectee 334 cases were infected during the pre-symptomatic stage of the infector. Separately, a modelling paper 335 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. Wölfel et al. [50] provides important data on a cohort of nine 'mild' cases which were serially tested 339 using sputum, swabs (throat and nasopharyngeal), urine and faecal samples over time. Importantly, 340 the virus was isolated, and inferences on viral replication could be made. Viral Isolation, and insights 341 into viral replication, improve inference around viral dynamics and transmission risk. The study 342

suggested high viral loads shortly after symptom onset, which declined thereafter over time. Positive 343 cultures were found from day 3-8 post-symptom onset ( Figure 5 ), and the minimum 5% isolation 344 success was achieved up to 9.8 (95% CI: 8.5-21.8) days post onset from throat and lung samples but 345 not faeces, blood or urine. 346 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. Inferring infectiousness was challenging given the heterogeneity of evidence available. Virological 348 diagnostic studies provide robust time series of infection, however, is limited by inferring the 349 relationship between PCR diagnostics and infectiousness. These data can also be affected by 350 sampling procedure and sample sites (e.g. upper respiratory, lower respiratory, faeces, urine, blood). 351

We have excluded RT-PCR durations based on faecal sampling due to the uncertainty whether these 352 data pertain to transmission potential ([50]; see below). Virological studies where culturing has 353 taken place, and where viral replication can be inferred would also be considered superior data to 354 infer infectious period, relative to estimates of viral load alone. We tested the hypothesis that severity of symptoms had an effect on symptomatic infectious 367 duration using a meta-regression approach. There was a trend towards studies that included severe 368 cases tended to have longer duration (estimated to be 4.0 days longer), but the effect was not 369 significant. Some studies have reported an association between duration of infectiousness and 370 incubation time (exposure to symptom on-set; mean 7.2 days), and short serial interval (mean 6.5 375 days; median 1.9 days; time from onset in primary to onset in secondary case). 376

Contact tracing studies provided robust evidence of transmission events, and therefore 377 infectiousness, but can be limited by the inferred timing of events, and symptoms experienced, due 378 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. symptoms during the presumed asymptomatic infectious period, which included "feeling warm" and 385 "feeling cold". However, the patient only "recognized getting sick" after she returned to China on 386 day four after the presumed exposure event. 387

Modelling parameters provide information on how COVID-19 data are being used and interpreted in 388 the research community, given the limited data available. Posterior estimates also provide 389 information on the parameter space at which infectious period central tendency reside, given other 390 parameters and assumptions in the model. Models used highly varied approaches to modelling 391 infectious period, which in turn resulted in highly variable parameter estimates used to inform the 392 studies. 393

There are few data for the precise definition of the asymptomatic infectious period (T2) parameter. 395 Some reported asymptomatic cases can actually be pre-symptomatic, when cases are subject to 396 follow-up (e.g.

[66]; see discussion above). However, Hu et al.

[7] do provide the data for 397 asymptomatic cases [that remain asymptomatic] across their presumed infectious period. Therefore, 398 in the first instance a parameter mimicking their data is probably the best available data. Note, there 399 is a large variation in this data parameter, and a gamma distribution of a shape alpha 3, beta 2, mean 400 6, may be appropriate for the initial model runs. Despite these being the primary informative data, 401 caution is required, given the uncertainty around the relationship between RT-PCR results and 402 infectiousness. Overall, an informed central tendency of ~6 days, with very low probability draws for 403 durations >20 days for the T2 parameter may be considered given the current state of knowledge. 404

The pre-symptomatic period is sometimes referred to as 'preclinical infectious' period (parameter 405 T3). This has been estimated from several papers, and the central tendency of these estimates vary 406 from <1 -4 days, cautiously approximating to 2 days, on average. also be measured as the difference between incubation and latent period, or the difference between 410 serial interval and incubation period.

[12] The relative consistency around the duration of this period 411 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. infectiousness was apparent on average 2.5 days prior to symptoms, reached a peak in risk at 0.6 440 days before symptoms, and decline up until 7 days after onset (9.5 days total infectious period). The 441

proportion of transmission before symptom onset (area under the curve) was estimated as 44% 442 (95% CI, 25-69%), based on inferences on incubation period. The authors suggest their data 443 supported the view that transmission risk decline substantially after 7 days post-symptoms onset. 444 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. 

Overall, the studies included were of good quality, though due to the rapid need for information 459 from the global research community many papers are pre-prints that have yet to be reviewed (at 460 time of writing). Many papers were limited in terms of sample sizes, with several papers being case 461 studies of one patient or single cluster outbreaks. There was a diversity of methods employed to 462 infer dynamics of infectiousness across studies, and therefore the evidential base was variable. Some 463 issues around nomenclature were noted, including definitions of asymptomatic, infectious period, 464 latent, and incubation period. It is possible the same data may have been used across different 465 studies, especially where publicly available data were used. 466

There was significant heterogeneity across study findings, and this was related to diversity of clinical 467 findings and methods employed. The meta-analysis employed for one parameter (T5) Another limitation is that a systematic review was not undertaken to inform this research, hence 477 there is a possibility that some relevant studies were overlooked. However, comprehensive search 478 strategies were conducted by two independent research groups to inform this research, hence 479 limiting the potential for missing key studies. 480

There are few data to inform asymptomatic infectious period (T2 parameter). One study provide 482 data that suggest a median period of 4-9.5 days, however, given the viral dynamics, this distribution is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . 586 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . Infection. Ann Acad Med Singapore 2020;49:1-9.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . https://doi.org/10.1101/2020.04.25.20079889 doi: medRxiv preprint CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . https://doi.org/10.1101/2020.04.25.20079889 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 30, 2020. . . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. coupled with a probability function to capture the maximal probability if exposed to a primary case. 804 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted April 30, 2020. . https://doi.org/10.1101/2020.04.25.20079889 doi: medRxiv preprint

CoV-2: an observational cohort study. The Lancet Infectious Diseases

Comparisons of nucleic acid conversion time of SARS-664

CoV-2 of different samples in ICU and non-ICU patients

Clinical Infectious Diseases: An Official Publication of the 667 Infectious Diseases Society of America

Asymptomatic and Presymptomatic SARS-669

CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility-King County

Quantifying SARS-CoV-2 transmission suggests 672 epidemic control with digital contact tracing

Suppression of COVID-19 outbreak in the 674 municipality of

The early phase of the COVID-19 outbreak in 676

Epidemiological and clinical characteristics of COVID-19 in 678 adolescents and young adults

Study claiming new coronavirus can be transmitted by people 680 without symptoms was flawed

Failed detection of the full-length genome of SARS-CoV-2 682 by ultra-deep sequencing from the recovered and discharged patients retested viral PCR 683 positive

Prolonged presence of SARS-CoV-2 in feces of pediatric 685 patients during the convalescent phase

Conducting quantitative synthesis when 687 comparing medical interventions: AHRQ and the Effective Health Care Program

Clinical course and risk factors for mortality of adult 690 inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

A Case Series of children with 2019 novel coronavirus 696 infection: clinical and epidemiological features

COVID-19 in a Designated Infectious Diseases Hospital 698

Outside Hubei Province, China

Recurrence of positive SARS-CoV-2 RNA in COVID-19: A 700 case report

First case of Coronavirus Disease 2019 702 (COVID-19) pneumonia in Taiwan

A case of COVID-19 and pneumonia returning from 704

Macau in Taiwan: Clinical course and anti-SARS-CoV-2 IgG dynamic

Persistence and clearance of viral RNA in 2019 novel 707 coronavirus disease rehabilitation patients

Patients of COVID-19 may benefit from sustained 709 lopinavir-combined regimen and the increase of eosinophil may predict the outcome of 710 COVID-19 progression

Positive result of Sars-Cov-2 in sputum from a cured 712 patient with COVID-19

Clinical Characteristics on 25 Discharged Patients with 714 COVID-19 Virus Returning

Clinical progression of patients with COVID-19 in Shanghai

The first infant case of COVID-19 acquired from a 719 secondary transmission in Vietnam. The Lancet Child & Adolescent Health

Clinical and epidemiological features of 36 children with 721 coronavirus disease 2019 (COVID-19) in Zhejiang, China: an observational cohort study

The Lancet Infectious

Prolonged presence of SARS-CoV-2 viral RNA in faecal 724 samples

Fundamental principles of epidemic spread 726 highlight the immediate need for large-scale serological surveys to assess the stage of the 727 SARS-CoV-2 epidemic