Asymptomatic people with SARS-CoV-2 as unseen carriers of COVID-19: A systematic review and meta-analysis

The asymptomatic patients with SARS-CoV-2 can act as an unseen carrier for magnifying the transmission of COVID-19. This study was designed to appraise the burden of asymptomatic individuals and estimate their occurrence among different age groups and gender by reviewing the existing published data on asymptomatic people with COVID-19.

this study was designed to evaluate the burden of asymptomatic individuals and estimate their occurrence among different age groups and gender by summarizing the existing published data on asymptomatic people with COVID-19.

Methods
Search strategy and data sources Electronic databases such as PubMed, Embase and Web of Science (WoS) were used to search studies following the PRISMA guideline [28]. PRISMA checklist (Table S1, Supplementary information (SI) le) was strictly followed while conducting this study. The following Medical Subject Headings (MeSH) terms and key words were used to search the published articles: "Novel coronavirus 2019", or "2019 nCoV", or "COVID-19", or "Asymptomatic carriers of COVID-19", or "Asymptomatic infections with COVID-19" and "SARS-CoV-2", where search was limited to articles published from January 1, 2020 to April 31, 2020 in English language.
Selection criteria Authors (GS, PD, SB and DRJ) evaluated search results and determined the eligibility of studies independently and dissonance was resolved by discussion between the authors. Any discrepancy during the review of full articles was resolved with a vote of majority.

Eligibility criteria
We considered following inclusion criteria for this meta-analysis and review: Study populations included asymptomatic COVID-19 infected patients (Cases who were tested positive for SARS-CoV-2 nucleic acid test by RT-PCR but without symptoms while screening).
Study designs included case-control study, prospective/retrospective cohort study and case reports. The inclusion criteria included: The exclusion criteria set for this study were: Study without asymptomatic proportion of COVID-19 infected patients Review and opinion articles, published protocols, meta-analyses, editorials and cases published in other languages than English.

Study selection
The results of the initial search strategy were rst screened for relevant articles by examining title and abstract. Then, full texts of relevant articles were inspected for inclusion and exclusion criteria. Studies that reported asymptomatic proportion of COVID-19 infected patients were included for quantitative synthesis.

Data extraction
From the screened articles, items including rst author, type of study-the publishing institution, date of publication, site of study, sample size, (Table 1) age and gender of patients, asymptomatic and symptomatic cases were recorded (Supplementary information (SI) File 1). Outcome measurements The primary nding in this study was the prevalence of symptomless cases (event rates) among COVID-19 infected patients. This study had also other secondary outcome measures including prevalence of asymptomatic cases among different gender and age, and prevalence of true asymptomatic cases. In case of secondary outcomes, the denominator was adjusted to asymptomatic cases to extract prevalence of asymptomatic cases in different age and gender.

Risk of bias (quality) assessment in individual studies
The potential risk of bias of each included studies was assessed according to Newcastle-Ottawa Scale (NOS) for observational studies (case-control, cohort, and cross-sectional studies). Studies were graded out of 10 points (stars) as shown in Table S2 (SI le). The mean score of two reviewers (GS, SB) was considered for the decision. Any variation in individual scores was checked and resolved by an author (DRJ). Since we did not nd any standard cut off score, the studies scoring arbitrary value ≥ 6 were used for meta-analysis and review. This quality assessment was performed to assess the systematic error and external validity of studies and ultimately to reduce the risk of biases.

Publication bias assessment
The potential publication bias was assessed by plotting the standard error and precision with Logit event rate as funnel plots. The absence of publication bias among included studies was also con rmed by Begg's (Begg and Mazumdar rank Correlation) and Egger's test (regression intercept) considering p-value > 0.05 indicating no publication bias.

Data analysis
All the statistical analysis were done using (Comprehensive Meta-Analysis version 3) statistical software (https://www.meta-analysis.com/). Percentages were calculated to describe the distributions of categorical variables. The prevalence of symptomless cases of COVID-19 infection was expressed as proportion and 95% con dence interval, using the random effects model, and was presented as Forest plot. Cochran Q test and inconsistency index (I 2 ) were used to detect heterogeneity among studies, with a p-value < 0.05 indicating signi cant heterogeneity. I 2 values of < 25%, 25-75% and > 75% indicate low, moderate and high heterogeneity respectively [29].

Subgroup analysis
Subgroup analysis was performed according to targeted patient age groups (children, adult and elderly), gender groups (male and female) and clinical outcome of asymptomatic cases (pre-symptomatic and true asymptomatic). Individuals of age less than 18 years were considered as children as recognized by the United Nations Convention on the Rights of the Child [30]. Asymptomatic cases not showing any typical symptoms associated with COVID-19 during the time of hospitalization were categorized into "True asymptomatic cases" and cases not showing any symptom during incubation period for up to 14 days but becoming symptomatic later were categorized into "pre-symptomatic cases". Studies that have reported age, gender and true cases of asymptomatic COVID-19 patients were only included in the subgroup analysis of respective age, gender, and clinical outcome of asymptomatic cases.

Study characteristics
This systematic review included 16 studies that were published between January and April, 2020.
Majority of the studies were from China (n = 11), two from Japan and three from the USA, including a total of 2,788 SARS-CoV-2 infected patients (Table 1). This review integrated cross sectional and observational cohort study.

Publication bias
The funnel plot (SI, Fig. 4) showed symmetry, demonstrating the absence of publication bias among the included studies. The Begg's (Begg and Mazumdar rank Correlation) and Egger's (regression intercept) tests also con rmed that there was no evidence of publication bias among the included studies for  Table S4).

Discussion
The pooled prevalence of symptomless cases of COVID-19 was 48.2% in this meta-analysis of 2,788 infected patients. In a case series of 78 patients from 26 transmission cluster cases in Wuhan, China, similar estimate (42.3%) of COVID-19 infected patients were asymptomatic carriers [26]. However, considering the basic reproduction number (R 0 ) 2.5, Centers for Disease Control and Prevention (CDC) has estimated slightly low (35%) asymptomatic SARS-CoV-2 infection rate [31]. This indicates that the large numbers of asymptomatic cases of COVID-19 are prevailing in the community, seeding potential outbreak which requires vigilant control strategies to prevent the next episode of outbreak. Thus, use of masks and widespread testing for identi cation and quarantine of infected asymptomatic individuals are essential to combat the rapid spread of COVID-19 pandemic locally and globally [32].
Although SARS-CoV, SARS-COV-2 and MERS belong to the same genus Betacoronavirus, differences in disease transmission and clinical features have been reported. SARS and MERS were mainly associated with nosocomial spread whereas SARS-CoV-2 is widely disseminated in the community [33]. The prevalence of symptomless infection was comparatively less in MERS and SARS, which was about 9.8% [34] and 13% [35] respectively. The reason for higher asymptomatic infection of COVID-19 could be related to the high replication e ciency of SARS-CoV-2, which replicate 3 times more rapidly compared to SARS-CoV. Thus SARS-CoV-2 can rapidly disseminate to the pharynx, from where it can be shed prior to the activation of innate immune response and production of symptoms [36]. Another contributing factor could be due to the less signi cant induction of host interferon and proin ammatory response, which also distinguishes it from SARS-CoV [36]. The proportion of symptomless cases of COVID-19 has gradually increased since its rst outbreak in Wuhan, Hubei province, China [37]. One of the reasons behind this may be because of decreased toxicity of SARS-CoV-2 in the process of long chain of transmission [38].
The communicable period (interval from the rst day of positive nucleic acid test to the rst day of continuous negative test) of asymptomatic case of COVID-19 could be as long as a month [39]. Thus, asymptomatic patient may carry virus for a long period of time and potentially spread it to others silently. Potential SARS-CoV-2 transmission from asymptomatic individuals had been documented in many studies [23][24][25]39]. In a study for assessing the viral load of SARS-CoV-2 in upper respiratory specimens of infected COVID-19 patients, high viral load was obtained among asymptomatic, pre-symptomatic and symptomatic patients, suggesting a high potential of transmission regardless of symptoms [40]. The biologic evidence for this is aided by a study conducted in a prolonged care facility where infectious SARS-CoV-2 was cultured from upper respiratory tract specimens of pre-symptomatic and asymptomatic patients even before six days to the development COVID-19 associated symptoms [41].
Among the total SARS-CoV-2 infected patients included in this study, 54.4% were male and 45.6% were female. This implies that the susceptibility of SARS-CoV-2 infection was higher in male compared to female. This might be explained by the presence of immune related gene on the X chromosome and sex hormones by which both immune responses: innate and adaptive can be in uenced [42][43][44]. Females mount stronger both innate and adaptive immune responses compared to males (42) and therefore might be less susceptible to SARS-CoV-2 infection. Another contributing factor could be a higher likelihood of exposure to the virus in males due to occupational risk and more outdoor activities. Smoking behavior had also been linked to increased incidence of SARS-CoV-2 infection among males [45]. The gene expression of angiotensin -converting enzyme 2 (ACE2), which is a cell receptor for SARS-CoV-2, is signi cantly elevated in smokers [46]. Since smoking behavior is comparatively higher in males than in females in general, increased incidence of SARS-CoV-2 could have been observed in males.
Subgroup analysis of asymptomatic patients revealed that the children are the most predominant asymptomatic carrier of SARS-CoV-2 (49.6%), followed by adults (30.3%) and elderly (16.9%). The increased incidence of symptomless cases of COVID-19 in children could be related to both exposure and host factors. Children's immune system is not well developed, and it is speculated that maturity and binding ability of ACE2 in children is lower [47]. Additionally, children often experience many viral infections. Hence, it is possible that repeated viral exposure aids the immune system to respond against SARS-CoV-2 [48]. Meanwhile, elderly populations have a weakened immune system [49,50] and less likely to be an asymptomatic carrier. In contrast, adult population most likely has stronger immune system to tackle with the infection and can remain as asymptomatic carrier. However, an in-depth mechanism for difference in asymptomatic manifestation among these three age groups (children, adult, and elderly) is yet to be explored.
The pooled prevalence of asymptomatic SARS-CoV-2 infection was observed to be higher in females (55.5%) than in males (44.5%) from subgroup analysis. Similarly, more women (66.7%) were observed to be asymptomatic cases in a case series done among 78 close contacts of COVID-19 patients in Wuhan, China [26]. Although there is no signi cant subgroup difference according to gender, higher prevalence of symptomless cases in female could be owing to less exposure and host factors in female. In comparison to male, female exhibit stronger innate, cellular, and humoral immune responses due to increased activation effects of female sex hormones and the presence of immune response genes on sex chromosomes [51]. These enhanced innate and adaptive immune responses could have contributed to the increased asymptomatic SARS-CoV-2 infection in females.
Asymptomatic COVID-19 infected patients (cases who were tested positive for SARS-CoV-2 nucleic acid test by RT-PCR but not showing any typical symptoms of SARS-CoV-2 infection while screening) were included in this meta-analysis. The reported prevalence of symptomless SARS-CoV-2 in this study consisted prevalence of both true asymptomatic patients (cases not showing any signs and symptoms at all during the period of hospitalization) and pre-symptomatic patients (cases not showing any symptom during incubation period for up to 14 days but becoming symptomatic later). Among total asymptomatic SARS-CoV-2 cases included in the subgroup analysis, 39% were true asymptomatic cases and 15.3% were pre-symptomatic cases. After 3 days of RT-PCR nucleic acid testing positive, pre-symptomatic cases showed mild COVID-19 related symptoms including fever, malaise, and cough [52,53].
A recent review stated that, as the surveillance and contacts testing of the MERS progressed overtime, the rate of asymptomatic MERS infected patients increased up to 28.6% [34]. This increase in the asymptomatic infection of MERS was inversely proportionate to the case fatality rate. This clearly demonstrates the importance of mass surveillance and contact tracing in the detection of asymptomatic COVID-19 infected patients, reduction of fatality rate and control of similar respiratory infection of COVID-19. Transmission of COVID-19 through asymptomatic persons is the Achilles' heel of COVID-19 control and prevention strategies. Hence, mass testing of SARS-CoV-2 among asymptomatic persons should be broaden especially among amass living conditions like prisons, inpatients of hospitals, camps, nursing, and aged care facilities. To limit the rapid spread of this novel pandemic, screening and SARS-CoV-2 testing of asymptomatic carriers must be prioritized in the community level along with wakeful control strategies.
Several strengths of this study include comprehensive inclusion of 16 studies adding more precision to the estimation of the prevalence of asymptomatic SARS-CoV-2 infection. Subgroup analysis found that the children and female were the most likely asymptomatic carriers of SARS-CoV-2, highlighting the necessity of providing special attention to them for preventing and controlling this pandemic.
This study has few limitations. As asymptomatic patients in the community might have been unnoticed and missed from being detected, the pooled prevalence rate in this study might be under-reported. Availability of data and materials All data generated or analyzed during this study are included in this published article and its supplementary information le.

Figure 1
Page 20/21 PRISMA ow diagram of screening process for the selection of studies for meta-analysis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. PRISMAChecklist.docx Supplementaryinformation2.docx