Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: Update of a living systematic review and meta-analysis.

BACKGROUND: Debate about the level of asymptomatic Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection continues. The amount of evidence is increasing and study designs have changed over time. We updated a living systematic review to address 3 questions: (1) Among people who become infected with SARS-CoV-2, what proportion does not experience symptoms at all during their infection? (2) What is the infectiousness of asymptomatic and presymptomatic, compared with symptomatic, SARS-CoV-2 infection? (3) What proportion of SARS-CoV-2 transmission in a population is accounted for by people who are asymptomatic or presymptomatic? METHODS AND
FINDINGS: The protocol was first published on 1 April 2020 and last updated on 18 June 2021. We searched PubMed, Embase, bioRxiv, and medRxiv, aggregated in a database of SARS-CoV-2 literature, most recently on 6 July 2021. Studies of people with PCR-diagnosed SARS-CoV-2, which documented symptom status at the beginning and end of follow-up, or mathematical modelling studies were included. Studies restricted to people already diagnosed, of single individuals or families, or without sufficient follow-up were excluded. One reviewer extracted data and a second verified the extraction, with disagreement resolved by discussion or a third reviewer. Risk of bias in empirical studies was assessed with a bespoke checklist and modelling studies with a published checklist. All data syntheses were done using random effects models. Review question (1): We included 130 studies. Heterogeneity was high so we did not estimate a mean proportion of asymptomatic infections overall (interquartile range (IQR) 14% to 50%, prediction interval 2% to 90%), or in 84 studies based on screening of defined populations (IQR 20% to 65%, prediction interval 4% to 94%). In 46 studies based on contact or outbreak investigations, the summary proportion asymptomatic was 19% (95% confidence interval (CI) 15% to 25%, prediction interval 2% to 70%). (2) The secondary attack rate in contacts of people with asymptomatic infection compared with symptomatic infection was 0.32 (95% CI 0.16 to 0.64, prediction interval 0.11 to 0.95, 8 studies). (3) In 13 modelling studies fit to data, the proportion of all SARS-CoV-2 transmission from presymptomatic individuals was higher than from asymptomatic individuals. Limitations of the evidence include high heterogeneity and high risks of selection and information bias in studies that were not designed to measure persistently asymptomatic infection, and limited information about variants of concern or in people who have been vaccinated.
CONCLUSIONS: Based on studies published up to July 2021, most SARS-CoV-2 infections were not persistently asymptomatic, and asymptomatic infections were less infectious than symptomatic infections. Summary estimates from meta-analysis may be misleading when variability between studies is extreme and prediction intervals should be presented. Future studies should determine the asymptomatic proportion of SARS-CoV-2 infections caused by variants of concern and in people with immunity following vaccination or previous infection. Without prospective longitudinal studies with methods that minimise selection and measurement biases, further updates with the study types included in this living systematic review are unlikely to be able to provide a reliable summary estimate of the proportion of asymptomatic infections caused by SARS-CoV-2. REVIEW PROTOCOL: Open Science Framework (https://osf.io/9ewys/).

Introduction

There is ongoing debate about the true proportion of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection that remains asymptomatic [1]. A well-recognised source of overestimation arises when people without symptoms at the time of testing are reported as having asymptomatic infection, with such cross-sectional studies often reporting percentages of 80% or more [2,3]. These studies overestimate the proportion of persistently asymptomatic infection because they misclassify people with so-called presymptomatic infection, who will develop symptoms of Coronavirus Disease 2019 (COVID-19) if reassessed after an adequate follow-up period [1]. Other sources of bias can result in over- or underestimation of the proportion with persistent asymptomatic infections, even when participants are adequately followed up [1]. For example, studies that assess a limited range of symptoms could overestimate the proportion asymptomatic through misclassification if they do not ask participants about all possible symptoms. Since COVID-19 was first identified as a viral pneumonia, the spectrum of symptoms has grown to include gastrointestinal symptoms and disturbances of smell and taste [1]. On the other hand, selection bias would be expected to underestimate the proportion with asymptomatic SARS-CoV-2 if people with symptoms are more likely to be tested for SARS-CoV-2 infection than those without symptoms [4].

Accurate estimates of the proportions of true asymptomatic and presymptomatic infections are needed to determine the balance and range of control measures [5]. Recognition of asymptomatic and presymptomatic infections showed the importance of control measures such as physical distancing, active case-finding through testing of asymptomatic people [6], and the need for rapid quarantine [7] in the first waves. Since late 2020, vaccines have become available [8] and several SARS-CoV-2 variants of concern have spread internationally, with varying viral characteristics [9]. The number of published studies about SARS-CoV-2 is increasing continuously, and the types of published studies are also changing [10], including the designs of studies about asymptomatic infection. In systematic reviews of studies published to April 2021, reported point estimates from random effects meta-analysis models range from 17% to 41% [11-16]. Authors of these reviews typically report values of the I2 statistic >90% [17,18], but heterogeneity is often not explored in detail and prediction intervals, which give information about sampling error and variability between studies, are recommended but rarely reported [17,19,20]. In this fifth version of our living systematic review [21], we aimed to improve and understand the changing evidence over time for 3 review questions: (1) Among people who become infected with SARS-CoV-2, what proportion does not experience symptoms at all during their infection? (2) What is the infectiousness of people with asymptomatic and presymptomatic, compared with symptomatic SARS-CoV-2 infection? (3) What proportion of SARS-CoV-2 transmission is accounted for by people who are either asymptomatic throughout infection or presymptomatic?

Methods

We conducted an update of a living systematic review, a systematic review that provides an online summary of findings and is updated when relevant new evidence becomes available [22]. The protocol, which describes modifications for each version, was first published on 1 April 2020 and amended for this fifth version on 18 June 2021 (https://osf.io/9ewys/). Previous versions have been posted as preprints [21,23] and published as a peer-reviewed article [12]. We report our findings according to statements on preferred reporting items for systematic reviews and meta-analyses 2020 (S1 PRISMA Checklist) [24] and on synthesis without meta-analysis in systematic reviews (SWiM) [25]. Ethics committee review was not required for this review. Box 1 shows our definitions of symptoms, asymptomatic infection, and presymptomatic status.

Box 1. Definitions of symptoms and symptom status in a person with SARS-CoV-2 infections

Symptoms: symptoms that a person experiences and reports. We used the authors’ definitions. We searched included manuscripts for an explicit statement that the study participant did not report symptoms that they experienced. Some authors defined “asymptomatic” as an absence of self-reported symptoms. We did not include clinical signs observed or elicited on examination.

Asymptomatic infection: a person with laboratory-confirmed SARS-CoV-2 infection, who has no symptoms, according to the authors’ report, at the time of first clinical assessment and had no symptoms at the end of follow-up. The end of follow-up was defined as any of the following: virological cure, with 1 or more negative reverse transcription PCR (RT-PCR) test results; follow-up for 14 days or more after the last possible exposure to an index case; follow-up for 7 days or more after the first RT-PCR positive result.

Presymptomatic: a person with laboratory-confirmed SARS-CoV-2 infection, who has no symptoms, according to the authors’ report, at the time of first clinical assessment, but who developed symptoms by the end of follow-up. The end of follow-up was defined as any of the following: virological cure, with 1 or more negative RT-PCR test results; follow-up for 14 days or more after the last possible exposure to an index case; follow-up for 7 days or more after the first RT-PCR positive result.

Information sources and search

We conducted the first search on 25 March 2020 and updated it on 20 April 2020, 10 June 2020, 2 February 2021, and, for this update, 6 July 2021. We searched the COVID-19 living evidence database [26], which uses automated workflow processes to: (1) aggregate simultaneous daily searches of 4 electronic databases (Medline, PubMed, Ovid Embase, bioRxiv, and medRxiv), using medical subject headings and free-text keywords for SARS-CoV-2 infection and COVID-19; (2) deduplicate the records; (3) tag records that are preprints; and (4) allow searches of titles and abstracts using Boolean operators. We used the search function to identify studies of asymptomatic or presymptomatic SARS-CoV-2 infection using a search string of medical subject headings and free-text keywords (S1 Text). We also examined articles suggested by experts and the reference lists of retrieved studies. Reports were planned to be updated at 3 monthly intervals, with continuously updated searches.

Eligibility criteria

We included studies, in any language, of people with SARS-CoV-2 diagnosed by RT-PCR that documented follow-up and symptom status at the beginning and end of follow-up or investigated the contribution to SARS-CoV-2 transmission of asymptomatic or presymptomatic infection. We included contact tracing and outbreak investigations, cohort studies, case-control studies, and mathematical modelling studies. We amended eligibility criteria after the third version of the review [12] in 2 ways. First, we excluded studies that only reported the proportion of presymptomatic SARS-CoV-2 because the settings and methods of these studies were very different and their results were too heterogeneous to summarise [12]. Second, we aimed to reduce the risk of bias from studies with inclusion criteria based mainly on people with symptoms, which would systematically underestimate the proportion of people with asymptomatic infection. We therefore excluded the following study types: case series restricted to people already diagnosed and studies that did not report the number of people tested for SARS-CoV-2, from whom the study population was derived. We also excluded case reports and contact investigations of single individuals or families, and any study without sufficient follow-up (Box 1). Where data from the same study population were reported in multiple records, we extracted data from the most comprehensive report.

Study selection and data extraction

Reviewers, including crowdsourced-trained volunteers, worked in pairs to screen records using an application programming interface in the electronic data capture system (REDCap, Vanderbilt University, Nashville, Tennessee, United States of America). One reviewer applied eligibility criteria to select studies and a second reviewer verified all included and excluded studies. We reported the process in a flow diagram, adapted for living systematic reviews [27] (S1 Fig). The reviewers determined which of the 3 review questions each study addressed. One reviewer extracted data using a prepiloted extraction form in REDCap and a second reviewer verified the extracted data. For both study selection and data extraction, a third reviewer adjudicated on disagreements that could not be resolved by discussion. We contacted study authors for clarifications. The extracted variables included study design, country and/or region, study setting, population, age, sex, primary outcomes, and length of follow-up (full list of variables in S1 Appendix). We extracted raw numbers of individuals with an outcome of interest and relevant denominators from empirical studies. From statistical and mathematical modelling studies, we extracted proportions and 95% credibility intervals.

The primary outcomes for each review question were (1) the proportion of people with asymptomatic SARS-CoV-2 infection who did not experience symptoms at all during follow-up; (2) secondary attack rate from asymptomatic or presymptomatic index cases, compared with symptomatic cases; (3) model-estimated proportion of SARS-CoV-2 transmission accounted for by people who are asymptomatic or presymptomatic.

Risk of bias in included studies

After the third version of the review [12], we developed a new tool to assess the risk of bias because the study designs of included studies have changed. In previous versions, we used items from tools to assess case series [28] and the prevalence of mental health disorders [29]. The new tool assessed possible biases in studies of prevalence in general and COVID-19 in particular [4,30]. We developed signalling questions in the domains of selection (2 items), information (3 items), and selective reporting (1 item) biases (S2 Text). For mathematical modelling studies, we used a checklist for assessing relevance and credibility [31]. Two authors independently assessed the risk of bias, using a customised online tool. A third reviewer resolved disagreements.

Synthesis of the evidence

The data extracted and the code used to display and synthesise the results are publicly available: https://github.com/leonieheron/LSR_Asymp_v5. We used the metaprop, metabin, and metafor functions from the meta package (version 4.11–0) [32] and the ggplot2 package (version 3.3.5) in R (version 3.5.1). The 95% confidence intervals (CIs) for each study were estimated using the Clopper–Pearson method [33]. For review question 1, in contact or outbreak investigations, we subtracted the index cases from the total with SARS-CoV-2 infection, because these people were likely to have been identified because of their symptoms and their inclusion might lead to underestimation of the asymptomatic proportion [16]. For all meta-analyses, we used stratified random effects models. Where a meta-analysis was not done, we present the interquartile range (IQR) and describe heterogeneity visually in forest plots, ordered by study sample size [25]. We calculated the I2 statistic, which is the approximate proportion of between-study variability that is due to heterogeneity other than chance, and τ2, the between-study variance. We calculated 95% prediction intervals for all summary estimates, to give a likely range of proportions that would have been obtained in hypothetical studies conducted in similar settings [17,19,20]. We did subgroup analyses for prespecified characteristics of study design, setting and risk of bias, and post hoc analyses according to age group and geographic region, comparing groups using a χ2 test. We used meta-regression for post hoc analyses examining associations with study size and publication date. To compare our findings with other studies, we extracted the raw data from 5 systematic reviews [11,13-16] and calculated prediction intervals [17]. For review question 2, as a measure of infectiousness, we calculated the secondary attack rate as the number of SARS-CoV-2-infected contacts as a proportion of all close contacts ascertained. For each included study, we compared the secondary attack rate from asymptomatic or presymptomatic index cases with that from symptomatic cases in the same study. If there were no events in a group, we added 0.5 to each cell in the 2 × 2 table. We did not account for potential clustering of contacts because the included studies did not report the number and size of infection clusters consistently. We used the Hartung–Knapp method for random effects meta-analysis to estimate a summary risk ratio (with 95% CI) [34]. For review question 3, we reported the findings descriptively because of large differences between study settings, methods, and results. We did not construct funnel plots to examine bias across studies because their utility in studies reporting on proportions is not clear.

Results

The searches for studies about asymptomatic or presymptomatic SARS-CoV-2, on 25 March, 20 April, and 10 June 2020 and 2 February and 6 July 2021 resulted in 89, 230, 688, 4,213, and 3,018 records for screening, respectively (S1 Fig). Owing to changes in eligible study designs, this update excludes 67 articles from earlier versions (S1 Table). We included a total of 146 studies addressing 1 or more review questions; 130 empirical studies that estimate the proportion of people with asymptomatic SARS-CoV-2 (summarised in Table 1 and S2 Table) [35-164], 8 studies reporting on secondary attack rates [35,81,131,142,165-168], and 13 mathematical modelling studies [7,165,169-179]. At the time of the search on 6 July 2021, 5 records were preprints. We checked the publication status on 14 March and all remained as preprints [61,62,88,169,171]. The review period from 1 January 2021 onwards includes 52 publications, 3 of which collected data during the period when the alpha variant of concern [85,125,133] had been described and vaccines were being introduced.

Table 1

Summary of characteristics of studies reporting on proportion of asymptomatic SARS-CoV-2 infections (review question 1).

	Study design and setting					All studies
	Contact investigation	Outbreak investigation	Screening of defined population
			Community	Institutional	Occupational
Total studiesa, n	13	33	23	43	18	130
Publication date
January 2020–June 2020	5	9	3	3	4	24
July 2020–December 2020	5	17	11	17	6	56
January 2021 onwards	3	7	9	23	8	50
Region b
Africa	0	2	2	1	1	6
Americas	5	10	4	19	7	45
Southeast Asia	0	3	2	2	1	7
Europe	2	13	7	18	5	45
Eastern Mediterranean	0	0	3	2	2	8
Western Pacific	6	6	5	1	2	19
Follow-up method c
14 days after last possible exposure	7	11	2	3	4	27
≥7 days after diagnosis	11	27	19	35	16	108
Until negative RT-PCR result	2	4	9	12	6	33
Two or more follow-up methods	8	21	17	28	14	88
Age range of study participants
Children (<18 years)	1	1	0	3	0	5
Adults (18–65 years)	3	10	9	16	14	52
Older adults (>65 years)	0	7	0	6	0	13
All ages	7	14	10	15	2	48
No information about age	2	1	4	3	2	12
Total with SARS-CoV-2 infection, n	1,076	4,910	10,652	8,921	2,867	28,426
Asymptomatic SARS-CoV-2 infections	264	1,409	6,007	3,658	585	11,923
Sex of asymptomatic casesd
Male	1	33	1,419	30	16	1,499
Female	0	32	589	313	26	960

RT-PCR, reverse transcription PCR; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

aS2 Table reports the characteristics of each study included.

bWorld Health Organization regions.

cStudies could have more than 1 method of follow-up (S2 Table).

dNinety-nine studies did not report the gender of asymptomatic cases.

Proportion of people with asymptomatic SARS-CoV-2 infection

The 130 studies reported on 28,426 people with SARS-CoV-2 infection (11,923 defined as having asymptomatic infection) in 42 countries [35-164] (Table 1 and S3 Table). Among all 130 included studies, 88 studies used more than 1 method of follow-up to ascertain asymptomatic status (Table 1 and S2 Table). Only 22 of 130 studies reported the median or mean age [38,47,70,76,77,83,85,95,99,119-121,124,126,128,133,134,139,143,146,152,164] and only 5 studies included children only [65,67,110,115,118]. Only 31 studies reported the sex of people with asymptomatic SARS-CoV-2 (Table 1 and S2 Table) [38,47,51,53,70,71,75,76,83,85,95,99,107,119-122,124,126,128,133,134,139,143,146,147,150,153,158,162,164]. The types of included studies changed across the 5 versions of the review. In the first version [23], 6 of 9 studies were contact tracing investigations of single-family clusters. In this version, 2 main types of study design generated the study populations of people with SARS-CoV-2: contact tracing or outbreak investigation methods were used to identify and test potentially infected contacts (46 studies, referred to as contact and outbreak investigations); and studies that involved screening of a defined group of people in settings in the community, institutions, such as long-term care facilities, or occupational groups (84 studies, referred to as screening studies).

Between-study heterogeneity was considerable, so we did not estimate a mean proportion of asymptomatic infections overall, or for screening studies (Fig 1). The IQR of estimates for all 130 included studies (141 clusters) was 14% to 50% (prediction interval 2% to 90%). In 46 studies enrolling people found through contact or outbreak investigations, for example, in long-term care facilities, in aeroplanes, or on cruise ships, we estimated a summary estimate for the proportion asymptomatic (19%, 95% CI 15% to 25%, prediction interval 2% to 70%, IQR 8% to 37%). The estimated proportions of asymptomatic SARS-CoV-2 infections were similar in studies of contact investigations (16%, 95% CI 9% to 27%, IQR 8% to 38%, 13 studies) and outbreak investigations (20%, 95% CI 15% to 28%, IQR 8% to 38%, 33 studies) (S2 Fig).

Fig 1

Forest plot of proportion of people with asymptomatic SARS-CoV-2 infection, stratified by study design.

In contact and outbreak investigations, the summary estimate for meta-analysis was 19% (15%–25%) and the IQR was 8%–37%. In screening studies, the IQR was 20%–65%. Across all studies, the IQR was 14%–50%. The x-axis displays proportions. Where more than 1 cluster was reported, clusters are annotated with [cluster identity]. The IQR is given below the individual study estimates. The red bar shows the prediction interval. CI, confidence interval; IQR, interquartile range; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

In 84 screening studies, the IQR for estimates from individual studies was 20% to 65% and the prediction interval was 4% to 94% [41-48,50-54,56,58-63,67,69-71,73,74,79,83,86,88-97,99-101,103-105,107,109,110,114-117,119,120,122,124-126,128,130,133-138,143-145,147,150-156,158-164]. The ranges of estimates were similar in 3 settings in which screening studies were conducted; people in a community setting (23 studies, IQR 20% to 53%, prediction interval 2% to 96%), institutional settings such as nursing homes (43 studies, IQR 26% to 67%, prediction interval 5% to 93%), and occupational settings such as among groups of healthcare workers (18 studies, IQR 17% to 64%, prediction interval 3% to 95%) (S2 Fig).

Three studies had data collection periods from 1 January 2021. In 2 nursing home outbreaks, 3/4 SARS-CoV-2 infections in partially vaccinated residents [133] and 13/14 infections (with the alpha variant) [85] in full vaccinated residents were asymptomatic. One study among healthcare workers did not report on symptom status according to vaccination or variant of concern but found that 76/155 (49%) with reinfection compared with 273/1,704 (17%) with primary infection were asymptomatic [125].

Risk of bias in individual studies

There were risks of bias in all types of empirical studies (S3 Fig). In prespecified subgroup analyses according to risk of bias domains (S4 Table), statistical heterogeneity remained very high (I2 ≥ 84%) and the prediction intervals remained wide. In screening studies, the summary proportion in was lower in studies judged to be at low risk of information bias in the assessment of symptoms (29%, 95% CI 20% to 42%) than in studies at unclear or high risk of bias (47%, 95% CI 37% to 57%) (p = 0.03, test for subgroup differences). For all other domains, estimates of the proportion asymptomatic were not associated with the assessment of the risk of bias.

Factors associated with proportion of asymptomatic SARS-CoV-2

Table 2 shows analyses for factors prespecified in the protocol and post hoc (Table 2). Study design and setting was prespecified and explained 16% of the variance in heterogeneity; estimates ranged from 16% (8% to 29%) for contact investigations to 45% (35% to 56%) for screening studies in institutional settings (S2 Fig). The date of publication was associated with the estimate of the proportion asymptomatic (S4 Fig), increasing in more recent publications (p 0.03), although this only explained 4% of the variance in heterogeneity. There was some evidence of variability in different world regions (p = 0.06), explaining 9% of the heterogeneity. Sample size and age range of the study participants did not appear to influence the estimated asymptomatic proportion. In 3 systematic reviews that we reanalysed, prediction intervals were: 1% to 83% (241 studies [14]); 4% to 97% (95 studies [15]); and 3% to 89% (170 studies [16]). I2 values were between 94% and 99% (S2 Appendix).

Table 2

Summary of findings of subgroup and meta-regression analyses of factors associated with the proportion of asymptomatic SARS-CoV-2 infections.

Variable	Clustersa	Proportion at the reference value (95% CI)	τ 2 b	p-Value (subgroup difference/intercept)	Heterogeneity variance explainedc
Reference	141	0.32 (0.27–0.38)	2.19	-	-
Study design and setting d
Contact investigation	13	0.16 (0.08; 0.29)	1.87	<0.001	16%
Outbreak investigation	40	0.20 (0.14; 0.28)
Screening: community	24	0.39 (0.26; 0.53)
Screening: institutional	46	0.45 (0.35; 0.56)
Screening: occupational	18	0.41 (0.26; 0.59)
Age group e
All ages	57	0.28 (0.21–0.37)	2.11	0.38	4%
Adults (18–65 years)	52	0.36 (0.27–0.47)
Older adults (>65 years) only	14	0.25 (0.13–0.44)
Children (<18 years) only	5	0.27 (0.09–0.58)
Not reported	13	0.45 (0.27–0.66)
Geographic region e , f
Americas	47	0.37 (0.27–0.47)	2.00	0.06	9%
Europe	52	0.24 (0.17–0.33)
Western Pacific	20	0.35 (0.21–0.51)
Southeast Asia	8	0.22 (0.09–0.43)
East Mediterranean	8	0.59 (0.33–0.81)
Africa	6	0.47 (0.22–0.74)
Selection bias a , d
Low risk	54	0.34 (0.25–0.44)	2.19	0.67	0%
Unclear/high risk	87	0.31 (0.25–0.39)
Information bias, assessment of symptoms defining status a , d
Low risk	33	0.25 (0.16–0.36)	2.15	0.14	2%
Unclear/high risk	108	0.35 (0.28–0.42)
Information bias, misclassification based on follow-up a , d
Low risk	107	0.32 (0.25–0.38)	2.19	0.65	0%
Unclear/high risk	34	0.35 (0.24–0.47)
Selective reporting bias a , d
Low risk	126	0.33 (0.27–0.40)	2.17	0.37	1%
Unclear/high risk	15	0.25 (0.13–0.43)
Sample size g
Proportion at 50	-	0.29 (0.14–0.51)	1.91	0.06	13%
Proportion at 120	-	0.31 (0.24–0.39)
Proportion at 200	-	0.32 (0.24–0.40)
Publication date
Reference (first date, 19 Feb 2020)	-	0.21 (0.13–0.32)	2.09	<0.001	4%
Coefficient	-	0.50 (0.50–0.50)

CI, confidence interval; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

aTotal number of studies, 130; independent within-study clusters counted individually;.

bCommon heterogeneity parameter estimated within each subgroup.

cFormula for proportion of heterogeneity variance explained, .

dPrespecified analysis in review protocol.

eSubgroup analysis not specified in review protocol.

fWorld Health Organization regions.

gPrevalence estimated using the meta-regression model for the approximate values of the median (n = 46), the mean (n = 202), and the third quartile (n = 126) of study sample sizes.

Infectiousness of people with asymptomatic or presymptomatic SARS-CoV-2

Eight studies provided data to calculate and to compare secondary attack rates by symptom status of the index case (Fig 2) [35,81,131,142,165-168]. Seven studies compared the secondary attack rate from asymptomatic with symptomatic index cases (summary risk ratio 0.32 (95% CI 0.16 to 0.64, prediction interval 0.11 to 0.95) [81,131,142,165-167]. One study compared asymptomatic with presymptomatic index cases (summary risk ratio 0.19, 95% CI 0.02 to 1.46) [35] and 4 studies compared presymptomatic with symptomatic index cases (summary risk ratio 1.00 (95% CI 0.37 to 2.71, prediction interval 0.11 to 9.12) [81,142,166,167]. The risk of information bias, specifically in symptom assessment, was judged to be high or unclear in 5 of the 8 studies included (S3 Fig).

Fig 2

Forest plot of the secondary attack rate of SARS-CoV-2 infections, comparing infections in contacts of asymptomatic and presymptomatic index cases with infections in contacts of symptomatic cases.

The RR is on a logarithmic scale. The diamonds show the summary estimate and its 95% CI. The red bar shows the prediction interval. CI, confidence interval; E, number of secondary transmission events; N, number of close contacts; RR, risk ratio; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2; Symp., symptomatic individuals.

Contribution of asymptomatic and presymptomatic infection to SARS-CoV-2 transmission

We included 13 mathematical modelling studies (Fig 3 and S5 Table) [7,165,169-179]. The models in 9 studies were informed by analyses of data from contact investigations in China, South Korea, Singapore, and from an outbreak on the Diamond Princess cruise ship, using data to estimate the serial interval or generation time [7,165,170,172,173,176-179]. In 3 studies, the authors did not analyse any original data sources [169,174,175].

Fig 3

Forest plot of proportion of SARS-CoV-2 infection resulting from asymptomatic or presymptomatic transmission.

For studies that report outcomes in multiple settings, these are annotated in brackets. CI, confidence interval; SI, serial interval; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

Estimates of the contributions of both asymptomatic and presymptomatic infections SARS-CoV-2 transmission were very heterogeneous. For asymptomatic SARS-CoV-2 infection, 4 studies contributed 5 estimates [7,171,175,177]. In 3 studies, the estimates suggested a contribution to SARS-CoV-2 transmission of asymptomatic infection of less than 15%. One study estimated a higher proportion (69%, 95% CI 20% to 85%) with a wide credibility interval [177] (Fig 3). The estimates have large uncertainty intervals, and the disparate predictions result from differences in the proportion of asymptomatic infections and relative infectiousness of asymptomatic infection.

We included 12 studies providing 16 estimates of the contribution of presymptomatic transmission [7,165,169,170,172-179]. The models examined a range of epidemic settings and used different assumptions about the durations and distributions of infection parameters such as incubation period, generation time, and serial interval (S5 Table). In 7 studies, point estimates for the estimated contribution of presymptomatic infection to all SARS-CoV-2 transmission in at least 1 reported scenario were 40% or greater [7,169,170,172,173,175,176] (Fig 3). In 1 study that estimated a contribution of <1% [174], the model-fitted serial interval was longer than observed in empirical studies. The credibility of most modelling studies was limited by the absence of external validation and of uncertainty intervals for the estimates cited (S5 Fig). The estimates from studies that relied on data from different published sources that might not have been compatible were assessed as providing low quality evidence (S5 Table).

Discussion

Summary of main findings

(1) For all 130 included studies, the IQR for the proportion of asymptomatic SARS-CoV-2 was 14% to 50%, prediction interval 2% to 90%) and for 84 studies based on screening of defined populations, IQR 20% to 65% (prediction interval 4% to 94%). In 46 studies that identified participants through contact tracing of index cases and outbreak investigations, the summary proportion from meta-analysis was 19% (95% CI 15% to 25%, prediction interval 2% to 70%). (2) The risk ratio for the secondary attack rate from asymptomatic compared with symptomatic infections was 0.32 (95% CI 0.16 to 0.64, prediction interval 0.11 to 0.95) and for presymptomatic infections compared with symptomatic infection was 1.00 (95% CI 0.37 to 2.71, prediction interval 0.11 to 9.12). (3) In mathematical modelling studies, estimated proportions of all SARS-CoV-2 infections that result from transmission from asymptomatic individuals were mostly below 15%, and from presymptomatic individuals mostly higher than 40%. Evidence about asymptomatic infections caused by variants of concern, or with immunity following infection or vaccination, is limited.

Strengths and weaknesses of the living systematic review methods

A strength of the methodology of this review is the transparent reporting, with openly available data and changes over different versions reported in the protocol. Our inclusion criteria attempted to reduce risks of bias and we developed a new tool to address potential biases in the studies included in this review. In contact investigations, we subtracted index cases from the total number of people with SARS-CoV-2 to avoid underestimation of the proportion asymptomatic [16]. We examined heterogeneity in detail and, as a result of the wide prediction interval, we chose not to report an overall summary estimate [18,25]. A limitation of the methods for this living systematic review is that this update only includes published studies up to 6 July 2021. This covers the period when vaccines started to be rolled out and the alpha variant of concern became dominant in high-income countries. Although we made extensive efforts to comply with the planned 3 monthly updates, with weekly searches and a continuous process of screening, data extraction and risk of bias assessment, the pace of publications about SARS-CoV-2 exceeds the capacity of our crowd of reviewers [10,26]. Our decision to include preprints compensates for some of the delay because these articles appear sooner than peer-reviewed publications. In reviews of observational epidemiological studies, search terms are broad so the number of studies that needs to be screened is high, but the yield of included studies is low. The 4 databases that we searched are not comprehensive, but they cover the majority of publications and we do not believe that we have missed studies that would change our conclusions. We have also not considered the possible impact of false negative RT-PCR results, which might be more likely to occur in asymptomatic infections [180] and would underestimate the proportion of asymptomatic infections [181].

Comparison with other reviews and interpretation

The type of studies that provide estimates of the proportion of asymptomatic SARS-CoV-2 infections and heterogeneity between them has changed over the course of the pandemic. In our living systematic review, the prediction interval has widened from 23% to 37% in studies published up to 25 March 2020 [23], to 3% to 67% up to June 2020 [12], 2% to 89% up to 2 February 2021 [21] and remains at 2% to 90% up to 6 July. We found 3 systematic reviews, in which authors reported restriction to studies with adequate follow-up (S2 Appendix) [11,13,16]. In 2 of the reviews, authors also applied criteria to reduce the risks of selection bias, with summary estimates of 18% (95% CI 9% to 26%, I2 84%, 9 studies) [13] and 23% (95% CI 16% to 30%, I2 92%, 21 studies) [11]. In both reviews, with studies published up to mid-2020, many included studies used designs that we defined as contact or outbreak investigations (Fig 1, S2 Table, and S2 Fig). Sah and colleagues reviewed studies published up to April 2021 and their subgroup estimate from studies in long-term care facilities, which include many outbreak investigations, was 17.8%, 95% CI 9.7% to 30.3%, 15 studies [16]. The summary estimates from all these reviews are compatible with our estimate from 46 studies in similar settings (19%, 95% CI 15% to 25%, prediction interval 2% to 70%, I2 90%) (Fig 1).

It may not be possible to obtain a single summary estimate from published literature of the proportion of persistently asymptomatic SARS-CoV-2 infection. Estimates from meta-analysis might be precise, but are likely to be unreliable owing to unacceptably high levels of heterogeneity. In 3 large systematic reviews, overall estimates had narrow confidence intervals [14-16], but I2 values were 94% to 99% and prediction intervals, which show the extent of all between-study variability were not reported [17]. The prediction intervals that we calculated extended more or less from zero to 100% (S2 Appendix), making differences in estimates between these studies hard to interpret. We expected this update to our living systematic review to provide a more precise and less heterogeneous estimate of the proportion of people with asymptomatic SARS-CoV-2 than in the previous version [12]. In particular, we expected that studies that detect SARS-CoV-2 through screening of defined populations and follow up of those infected would be less affected by biases in study methodology [30] and would provide a more accurate estimate of persistently asymptomatic SARS-CoV-2, which should be influenced mainly by properties of the virus and the host response to infection [182]. Study design was the factor that explained the largest proportion of variability in this review (S2 Appendix). Information bias, resulting from the way in which asymptomatic status is determined, was the factor most strongly associated with the estimated proportion of asymptomatic infection in screening studies (Table 2). Studies based on contact and outbreak investigations might obtain more detailed data about symptoms, resulting in lower estimates of the proportion that is classified as asymptomatic. Symptom report also differs between different groups of study participants, even within the same study, and could also contribute to heterogeneity [183]. Age might play a role as children appear more likely than adults to have an asymptomatic course of infection, but age was poorly reported and could not be examined in detail (Tables 1 and 2).

The analysis of secondary attack rates in this update now provides strong evidence of lower infectiousness of people with asymptomatic than symptomatic infection (Fig 2). The difference in secondary attack rates between asymptomatic and symptomatic index cases in our meta-analysis is smaller and less biased than in systematic reviews that analyse groups of studies reporting asymptomatic index cases and of symptomatic cases separately [182,184]. In meta-analyses of 2 proportions, the direct comparison within studies reduces heterogeneity and is less biased [34]. Since SARS-CoV-2 can be transmitted a few days before the onset of symptoms [185], presymptomatic transmission likely contributes substantially to overall SARS-CoV-2 epidemics. If both the proportion and transmissibility of asymptomatic infection are relatively low, people with asymptomatic SARS-CoV-2 infection should account for a smaller proportion of overall transmission than presymptomatic individuals. This is consistent with the findings of modelling studies in our review, although the absence of descriptions of the epidemic context in many studies made it difficult to compare findings across studies.

Implications and unanswered questions

This living systematic review shows the challenges of synthesising evidence from observational epidemiological studies. Heterogeneity in systematic reviews of prevalence is a recognised challenge [34,186]. Methodological guidance to refrain from meta-analysis, and to report prediction intervals, when the variability between studies is extreme is often ignored in favour of summary estimates, which are easy to cite [18,20]. Part of the heterogeneity in our review arises from the fact that many studies were not designed to estimate the proportion of asymptomatic SARS-CoV-2 infection. The incomplete descriptions of inclusion criteria, response rates, follow-up, and of definitions of symptom status [1] made it difficult to assess the risks of bias and to investigate their contribution to between-study heterogeneity.

The finding that, in studies of contact and outbreak investigations, a substantial minority of people with SARS-CoV-2 infection remains asymptomatic throughout the course of infection, and that almost half of all transmission might occur before symptoms develop has already had implications for prevention. When SARS-CoV-2 community transmission levels are high, physical distancing measures and mask-wearing need to be sustained to prevent transmission from close contact with people with asymptomatic and presymptomatic infection. Integration of evidence from epidemiological, clinical, and laboratory studies will help to clarify the relative infectiousness of asymptomatic SARS-CoV-2. Studies using viral culture as well as RNA detection are needed since RT-PCR defined viral loads appear to be broadly similar in asymptomatic and symptomatic people [180,187].

Determining the viral dynamics and full clinical spectrum of infection with variants of concern is important. Variants classed as omicron differ substantially from all earlier SARS-CoV-2 variants, with high infectiousness and immune evasion [188], and viral characteristics and immunity could influence the occurrence of asymptomatic infection. Studies published in early 2022 are already reporting a wide range of estimates of asymptomatic omicron infection. In India, from the date of emergence of the omicron variant, 24 November 2021 to 4 January 2022, authors reported a high proportion of asymptomatic omicron variant infections (56.7% of 291) but did not report any follow-up and >80% of participants had been vaccinated [189]. In contrast, authors of a cohort study of an outbreak of omicron SARS-CoV-2 in Norway, found only 1 of 81 infections in a highly vaccinated group was asymptomatic after 10 days of follow-up [190]. There are increasing challenges for studies relying on routine health service or surveillance data; in many jurisdictions, indications for routine testing are being reduced, which will make selection biases more likely, and mandated quarantine and isolation periods for people with diagnosed SARS-CoV-2 infection are being reduced, which will increase information biases in the ascertainment of persistent asymptomatic status. Researchers need to design studies to address this specific research question for each variant of concern, taking into account vaccination status and prior infection. There are ongoing prospective studies that collect appropriate data [125], for which improved reporting could address the requirements for assessing asymptomatic infection status fully, but ongoing funding for these studies is not secure [191]. Without prospective longitudinal studies with methods that minimise selection and measurement biases, further updates to this living systematic review are unlikely to provide a reliable summary estimate of the proportion of asymptomatic infections caused by SARS-CoV-2.

(PDF)

Click here for additional data file.

Search strings.

(PDF)

Click here for additional data file.

Risk of bias tool.

(PDF)

Click here for additional data file.

Data extraction forms.

(PDF)

Click here for additional data file.

Analysis of other systematic reviews on asymptomatic SARS-CoV-2.

(PDF)

Click here for additional data file.

Studies included in version 3 and excluded in versions 4 and 5 of the living systematic review.

(PDF)

Click here for additional data file.

Characteristics of studies reporting on proportion of asymptomatic SARS-CoV-2 infections (review question 1 and question 2).

F, female; IQR, interquartile range; M, male; NR, not reported; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

(PDF)

Click here for additional data file.

Location of studies contributing data to review question 1.

(PDF)

Click here for additional data file.

Subgroup analysis according to risk of bias.

(PDF)

Click here for additional data file.

Characteristics of mathematical modelling studies.

CI, confidence interval; NPI, non-pharmaceutical intervention; SARS-CoV-2, Severe Acute Respiratory Syndrome Coronavirus 2.

(PDF)

Click here for additional data file.

Flowchart of identified, excluded, and included records as of 6 July 2021.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.

Risk of bias assessment of studies in question 1 and 2.

(PDF)

Click here for additional data file.

Forest plot of proportion of people with asymptomatic SARS-CoV-2 infection by date of publication.

(PDF)

Click here for additional data file.

Assessment of credibility of mathematical modelling studies.

NA, not applicable; NEI: not enough information; NR: not reported; PY: partially yes.

(PDF)

Click here for additional data file.

24 Jan 2022

Dear Dr Low,

Thank you for submitting your manuscript entitled "Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: a living systematic review and meta-analysis" for consideration at PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission out for external assessment.

However, before we can send your manuscript for assessment, we need you to complete your submission by providing the metadata that are required. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Jan 26 2022 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for full assessment.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Richard Turner, PhD

Senior Editor, PLOS Medicine

plosmedicine@plos.org

17 Feb 2022

Dear Dr. Low,

Thank you very much for submitting your manuscript "Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: a living systematic review and meta-analysis" (PMEDICINE-D-22-00260R1) for consideration at PLOS Medicine.

Your paper was discussed among the editors and sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to invite you to submit a revised version that addresses the reviewers' and editors' comments fully. You will recognize that we cannot make a decision about publication until we have seen the revised manuscript and your response, and we expect to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We hope to receive your revised manuscript by Mar 10 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

Please let me know if you have any questions, and we look forward to receiving your revised manuscript.

Sincerely,

Richard Turner, PhD

Senior editor, PLOS Medicine

rturner@plos.org

-----------------------------------------------------------

Requests from the editors:

As requested by one or more of our reviewers, please update the search to the most recent date possible.

After the abstract, please add a new and accessible "author summary" section in non-identical prose. You may find it helpful to consult one or two recent research papers published in PLOS Medicine to get a sense of the preferred style.

At line 61, we suggest "we conducted an update ..." or similar.

Please remove the information on funding and competing interests from the end of the main text. In the event of publication, this will appear in the article metadata, via entries in the submission form.

Noting one entry in table 2, please quote exact p values or "p<0.001" throughout, unless there are specific reasons to do otherwise.

Throughout the text, please remove spaces from within the square brackets for reference call-outs (e.g., "... more [2,3].").

In the reference list, please add "[preprint]" to all preprint citations, noting reference 33 and others.

Noting reference 10, please use the journal name abbreviation "PLoS Med."; and for reference 57 "PLoS ONE".

Noting reference 95 and others, please ensure that all references have full access details.

Thank you very much for including the PRISMA checklist. Please adapt this so that individual items are referred to by section (e.g., "Methods") and paragraph numbers, not by line or page numbers as these generally change in the event of publication.

Please remove the attached earlier version of the systematic review.

Comments from the reviewers:

*** Reviewer #1:

This is an excellent review of an important set of questions

The review is stated as being a "living" review, but the literature was only searched up to February 2021. So what is the definition of a living review, according to these investigators. A review of publications up to one year ago does not seem to fit with definitions of living reviews.

While it is understandable that the production of literature and resources to conduct the review process is limiting the ability to keep up to date, it would be good to think about restrictions that could be made to streamline this process and allow for more recent updating (eg dropping pre-prints, increasing minimal sample size for eligibility, restricting to particular study designs, excluding based on certain quality criteria). The WHO Covid-19 database is a one-stop shop that could be used instead of searching 4 separate databases

More updated information may also contribute information about the predominant variant in circulation at the time of the study.

The Comparison with other reviews is helpful and could be summarised graphically.

*** Reviewer #2:

This is a thoughtful and well-written update to a living systematic review and meta-analysis examining the prevalence and roles of asymptomatic/presymptomatic infection on the spread of COVID-19. To rely perhaps a bit too much on Greek myths, what began as a Herculean task now seems a Sisyphean one, given the vast numbers of papers and preprints on COVID-19. This update only goes through Feb 1, 2021, and thus misses the potential impact of immunity from infection and/or vaccination on the extent of symptomatic disease and transmission and similarly how these factors change by the dominant variant. I appreciate the authors reasoned approach to noting the massive heterogeneity in the studies reviewed and their caution in applying meta-analysis tools not appropriate for the massive variability in findings across studies.

A primary takeaway seems to be that these types of reviews may be overwhelmed by both the tsunami of reports and the rapidly changing landscape with widespread infection and vaccination and shifting variants, where each of these factors may contribute to differences in results for each of the three questions this study focuses on. As such, my main comment for the authors is to clarify how much they expect variants and immunity from both vaccination and infection to play a role and whether their study is focused just on 'wild-type' or is generally a question that can be answered across the multiple dimensions of immune history, dominant variant, and potentially other elements. The abstract and paper end by mentioning 'wild-type' SARS-CoV-2. Given the D614G mutation that was dominant even within the timeframe of this review/metaanalysis, was this really only a study of wild-type SARS-CoV-2 (as the last lines of the abstract and paper suggest)? It seems that no prospective study could be done to look at these questions any longer for that strain, but what about for other strains? I'd be interested in the authors thoughts on what kinds of studies will be needed now, given widespread population immune experience from vaccination & infection, and repeated replacement of dominant variants.

*** Reviewer #3:

This manuscript provides an update to the previous systematic review and meta-analysis of asymptomatic COVID-19 infections. Using 94 studies, the authors found that 13-45% of infections are asymptomatic, although the authors note high between-study heterogeneity. While the paper is well written, in my opinion it does not add a lot to the growing body of literature that has been published since the previous version.

My major concern with this article is that the database was last updated a year back on February 2, 2021. To justify a publication as a living systematic review and meta-analysis, the database needs to be updated. Additionally, as the authors mention in the introduction, several recent meta-analyses on asymptomaticity have been published, providing more up to do date systematic reviews and meta-analyses of the literature. For example:

Ma et al. performed a similar analysis with their database updated on Feb 4, 2021. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2787098

Number of studies included in the meta-analysis of asymptomatic percentage: 95

Sah et al. Last database update: April 2021

https://www.pnas.org/content/118/34/e2109229118#sec-3

Number of studies included in the meta-analysis of asymptomatic percentage: 170

I also have concerns about the analyses performed in the study which I highlight below:

Following my previous point, Table 1 indicates that only 13 studies published between January 2021 - June 2021 were included in the analysis which is probably because the database wasn't updated after February 2021. However, page 11 line 257-258 mentions "At the time of the latest search date,17 records were preprints, 14 of which had been published in peer-reviewed journals by 23 November 2021. Does this mean that another search was performed on 23 November 2021? If there was another search performed in November 2021, why weren't any studies published after June 2021 included in the analysis?

Definition of asymptomatic infection: The analyses include studies where COVID diagnosis was made based on antibody test rather than RT-PCR test. For example, the study Pirnay JP, 2020 observed four individuals with no symptoms, but only two of these were positive RT-qPCR. The other two were negative RT-qPCR but tested positive on antibody tests. The study itself notes "two individuals (M8 and M9), only the rapid antibody tests were positive, making it uncertain whether they actually had the disease. This manuscript assumes that these two individuals were asymptomatic. However, since the dates of exposure for these two individuals are uncertain, assuming that the positive antibody (but PCR negative) test is an indication of active infection is incorrect. As such, any study that used an antibody test to diagnose an asymptomatic infection should be excluded from the analysis.

Lack of symptom follow-up in studies included in the analysis: Some of the studies included in the analysis do not report an absence of symptoms at the end of follow-up, but despite this lack of data, these individuals are categorized as asymptomatic in this manuscript. For example, Bi Q, 2020 (reference 2, Table S2) mentions "At the time of the first clinical assessment …..and 17 (20%) of 87 had no symptoms." The study does not specify that these individuals remained symptom-free at the end of follow-up, and so cannot be categorized as asymptomatic. Such studies should be removed from the meta-analysis of asymptomaticity.

Bias arising due to over-representation of symptomatic cases: Since the studies included in the meta analysis were not designed to calculate asymptomatic percentage, there is an over representation of cases symptomatic cases in most studies, biasing the asymptomatic percentage towards a lower estimate. For example, Lee JY, 2020 investigated inpatients diagnosed with COVID-19 where 20% of the patients had severe disease. In such a setting, it is likely that the majority of patients were admitted because they were exhibiting symptoms and therefore asymptomatic individuals were likely under-represented. Additionally, surveillance studies would also suffer from this bias due to a higher willingness of symptomatic individuals to participate/provide consent. This issue has been highlighted and corrected by a recent meta-analysis (Sah et. al, 2021). Although performing a subgroup analysis by study design attempts to approach this issue, all study designs (except the ones where 100% of the population was sampled) will have this inherent selection bias. Without a correction, the estimates of asymptomaticity presented in this meta-analysis are unreliable.

Additional analysis: Performing a subgroup analysis of asymptomatic proportion by publication date is likely to be biased due to the publication efficiency of different journals (it may take anytime between 1 month to 6m+ to publish an article). A more informative analysis would look at the last date of data collection mentioned in the study.

*** Reviewer #4:

This paper is clearly written and well organized. I suggest a minor revision. Meanwhile, some information shall be updated in this article.

[see attachment]

*** Reviewer #5:

This is an interesting and useful systematic review and meta-analysis on the occurrence and transmission potential of asymptomatic and presymptomatic SARSCoV-2 infections. The methods including the search, study selection and data extraction, risk of bias assessment, synthesis of the evidence using meta and ggplot2 packages in R and random-effects models are mostly adequate. However, there are still a few major issues needing attention.

1) As the authors said in conclusion "this review does not provide a summary estimate of the proportion of asymptomatic SARS-CoV-2 across all study designs". With very wide interquartile ranges (IQR) in most of meta estimates which is largely due to extremely high heterogeneity between studies, it basically demonstrated that the current meta-analysis in this paper doesn't work or the way the evidence was synthesed may not be adequate or practical. With the overwhelming differences in attributes/factors in these 94 studies includes, it is extremely challenging and ambitious to do a meaningful meta analysis. Another substantial limitation in this study is that there is no information on Covid variants or vaccination status of paticipants, which may have a big impact of heterogeneity of the meta results and subsequently led to inconclusive results and findings. The sub group analyses showed some better and narrow IQRs which might indicate a direction of future work in meta-analysing these studies. One of useful features of this study is that authors identified quite a few difficulties, challenges and practical issues in conducting the meta-analysis on asymptomatic infections, however it would be better if the authors could go further (e.g., some relatively neat sub group meta analysis, or extra meta regression analysis) and to be constructive with more in depth and critical discussions to shed some lights in the tunnel.

2) Prediction interval. Prediction intervals were frequently used in the paper. However this measure is not standard in meta-analysis. I can see prediction intervals were calculated using the 3 previous meta-analyses results but not clear how they were calculated exactly. Basically, as this is a key element of this study, we need a detailed paragraph explaining what this prediction interval is for and what it means in practical terms and how it is relevant to the the summary estimate of meta-analysis. So far, these extremely wide prediction intervals are basically saying no to all the meta analysis results, but the quesion is, is it valid or appropriate doing so?

3) To go deeper in finding out the factors contributing to heterogeneity and wide IQR in meta results, I would suggest that authors to carry out some extra meta regression analyses to identify which factors have impact on meta results, e.g., time, age, region, ethnicity, and different study designs and settings so that can give some solid suggestions for future research on this topic.

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: review_letter0215.docx

Click here for additional data file.

18 Mar 2022

Submitted filename: Replytocomm_asymptomaticreview_upload.docx

Click here for additional data file.

8 Apr 2022

Dear Dr. Low,

Thank you very much for re-submitting your manuscript "Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: update of a living systematic review and meta-analysis" (PMEDICINE-D-22-00260R2) for consideration at PLOS Medicine.

I have discussed the paper with our academic editor and it was also seen again by all reviewers. I am pleased to tell you that, provided the remaining editorial and production issues are fully dealt with, we expect to be able to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We hope to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

Please let me know if you have any questions, and we look forward to receiving the revised manuscript.

Sincerely,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org

------------------------------------------------------------

Requests from Editors:

Please check the version numbers for consistency, e.g., "v4" at Github, "fifth update" at line 161 and "first three versions" in the acknowledgements.

At line 83 (abstract), should that be "0.95"?

In table 1, should that be "sex" rather than "gender"?

At line 270, please make that "8 studies" and use this general style throughout, although numbers should be spelt out at the start of sentences.

At line 396, please revisit "0.38 95% 0.16-0.64", which appears inconsistent with the numbers quoted in the abstract, for example.

Please revisit reference 26, which may contain a superfluous close bracket.

Please remove the asterisks from reference 50.

Please make that "JAMA" in reference 71 and any other instances; and "PLoS ONE" in reference 134.

Please make that "US" in reference 92; and "UK" in reference 191.

Please check the journal name in reference 109.

Should "MFM" be removed from reference 153?

Comments from Reviewers:

*** Reviewer #1:

Excellent review that I hope will be published soon.

I still have a concern about claiming this as a living review. It sets expectations from an end user perspective and as this will likely only be published a year after the end date of the last search.

But leave it to the authors if they think this is important.

*** Reviewer #2:

I am satisfied with the authors' responses and believe the manuscript is acceptable for publication.

*** Reviewer #3:

The authors have satisfactorily addressed my previous concerns. To facilitate replication of results in the future, my only remaining suggestion is to include study characteristics described in Table 1 to individual studies listed in Table S2, including the study design, number of index cases excluded (if any), study region and age range of study participants.

*** Reviewer #4:

The authors have updated the review with 40 additional studies. All questions were answered in detail and necessary changes were made carefully.

I suggested this article be accepted.

*** Reviewer #5:

Many thanks authors for their great effort to improve the manuscript. I am satisfied with the response and revision especially the meta regression analysis was nicely done. No further issues needing attention.

***

Any attachments provided with reviews can be seen via the following link:

[LINK]

12 Apr 2022

Submitted filename: Replytocomm_asymptomaticreview_minor_220412_upload.docx

Click here for additional data file.

13 Apr 2022

Dear Dr Low,

On behalf of my colleagues and the Academic Editor, Dr Ford, I am pleased to inform you that we have agreed to publish your manuscript "Occurrence and transmission potential of asymptomatic and presymptomatic SARS-CoV-2 infections: update of a living systematic review and meta-analysis" (PMEDICINE-D-22-00260R3) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper.

Sincerely,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org