Epidemiologic, Clinical, and Laboratory Findings of the COVID-19 in the current pandemic: Systematic Review and Meta-analysis

Yewei Xie University of North Carolina at Chapel Hill-China Project https://orcid.org/0000-0002-9280-5812 Zaisheng Wang University of North Carolina at Chapel Hill-China Project Huipeng Liao University of North Carolina at Chapel Hill-China Project Gifty Marley Nanjing Medical University Dan Wu London School of Hygiene and Tropical Medicine Weiming Tang (  weimingtangscience@gmail.com ) https://orcid.org/0000-0002-9026-707X


Background
The emergence of COVID-19 has made it the rst infectious disease pandemic in the 21st century. As of April 23rd, a total of 2,544,792 COVID-19 cases has been reported and con rmed with 175,694 deaths globally. While more than 30 countries had issued the highest level of response, the SARS-CoV-2 (pathogen of COVID- 19) continues to spread in different regions around the world [W1]. However, the key information on the virus epidemiology, clinical spectrum, and on the pathogen were delayed in response during the early outbreaks in many countries. To ll the research gaps mentioned above, this article systematically summarizes global ndings on the natural history, clinical spectrum, transmission patterns, laboratory ndings, CT results, and risk factors of the COVID-19.

Information sources, search and study selection
We searched for publications in epidemiology and clinic domains of the COVID-19 broadly as it was a novel, rapidly extended, infectious disease. Both the English and Chinese were used in searching as the COVID-19 was rstly reported in Wuhan, most of the real-time study were in Chinese. Then, we searched the following databases: CHKD v3.1 of the CNKI [in Chinese], PubMed, Web of Science, and medRxiv&bioRxiv by using search terms 'COVID-19, SARS2, SARS-CoV-2, 2019 nCoV' (Appendix 1). The publication date was restricted from 1st Dec 2019 to 23rd Apr 2020. Only the full-text available human studies were eligible for selection. Real-time data were obtained from health departments of multiple countries, global NGOs, and reputable media sources.

Data collection process
Three review authors (HL, YX, and ZW) reviewed the titles and abstracts of all searched records in the rst screening. The full-text articles were reviewed for further screening. The full-text records would submit to the reviewer (DW and WT) to consult and make nal decisions when there were disagreements. The Four authors (HL, WT, YX, and ZW) extracted data from the included studies independently. When there were disagreements, the data were handed to the full members of the review group for further discussion and resolved. At last, the con rmed data will be entered into Review Manager 5 (RevMan 5) and doublechecked by all review authors in the team.
The Cochrane Handbook for Systematic Reviews of Interventions suggested review authors collect missing data from investigators the missing data. It was because the missing data could lead to the bias during analyzing. However, using the imputation method to tackle the missing data problem could not reduce bias [W2]. Therefore, we only analyze data available to us if we could not collect the missing data from the investigators.
Dealing with missing data The Cochrane Handbook for Systematic Reviews of Interventions suggested review authors collect missing data from investigators the missing data. It was because the missing data could lead to the bias during analyzing. However, using the imputation method to tackle the missing data problem could not reduce bias [W2]. Therefore, we only analyze data available to us if we could not collect the missing data from the investigators.

Data synthesis, and assessment of heterogeneity and reporting biases
We reviewed all kinds of researches about the COVID-19 outbreak systematically from the beginning until the latest date before we submitted this paper. Because the objective of this study aimed to get a bird view of the COVID-19 outbreak, and updated the clinical spectrum and natural history of the COVID-19, we synthesis data narratively. Thus, to assess the heterogeneity and reporting bias of the included studies.

Results
We collected 5767 records from 4 databases (CNKD, Pubmed, Web of Science, and medVix&BioVix), and 20 records from other resources. 5410 records remained after removing 377 duplications. After two batches of screening, 78 records were included in this review.

Demographic characteristics
In a 55,924 COVID-19 patients study based in China, the majority of patients aged 30-69 (77.8%) with only 2.4% of the patients being 18 years and below. The median age of the patients was 51 (ranged 2 days-100 years old) [W3]. Similarly, in the United States, more than half of patients aged between 20-64 years (65%), and patients under 19 only accounted for 5% of all patients. Older aged patients were more prone to getting infected compared to the young [1]. The US's data also indicated that patients younger than 19 had milder COVID-19 illness, with almost no hospitalizations or deaths reported [W4]. Based on a worldwide data, the elderly (aged over 60) were at a high risk of developing into severe disease or death [W3-W5, 2]. About the gender ratio, the male to female ratio of con rmed cases was 1.06:1.00 in China [3]. However, in South Korea and Iceland, the male population had a higher incidence rate than the female population [1,2]. Males had double times the secondary attack rate than females [4].

Transmission stages
Globally, the transmission of the COVID-19 can be categorized into 4 temporal stages. The rst stage: people with exposure histories to Huanan Seafood Market (HSM) got infected [5]. Forty-one patients were found to be having SARS-like symptoms in December 2019, and the HSM was believed to be one of the origins of the virus. However, 13 of the 41 patients reported no prior exposure to the HSM thus indicating that the origin of the virus needed further investigation [6]. The second temporal stage is the transition from community transmissions to the outbreak in Wuhan [5]. The virus was mainly spread to multiple communities directly and indirectly by people with HSM exposure histories. The interpersonal transmissions and clustered transmissions formed community transmissions [5]. An early study showed that the proportion of patients with HSM exposure histories decreased from 55-8.6% within 22 days, indicating people who did not have exposure histories to the HSM became infected [7,8]. The third stage: the epidemic in China. At this stage, transmissions began to expand to communities outside Wuhan and the Hubei province as a whole [5]. On 26th Jan 2020, a study involving 62 COVID-19 patients outside Wuhan found that all the patients had been exposed to Wuhan, which demonstrated an established local transmission outside Wuhan [9]. The fourth temporal stage is the global pandemic. On 13th Jan 2020, the rst case outside China was reported in Thailand [W1]. On 30th Jan 2020, the WHO declared a Public Health Emergency of International Concern (PHEIC) [W1]. It subsequently took about 51 days for transmission to escalate from the rst reported case to the 10,000th reported case outside China. Globally, it took 16 days for the number of reported cases to increase from 10,000th cases to 100,000th cases, 21 days from 100,000th cases to 500,000th cases, only 6 days from 500,000th cases to 1,000,000th cases and 13 days from 1,000,000th cases to 2,000,000th cases [W1].

Transmission Routes
The main transmission form of this virus was by human-to-human spread, since only 1.18% patients among 1099 con rmed patients had history of direct contact with wild animals [10]. The vital transmission routes were through respiratory droplets and contact transmissions. There remains the possibility of aerosol transmission when exposed to high concentrations of aerosols for a long time in a relatively closed environment [W6]. Fecal-oral transmission and the mother-to-child transmission were also possible but lacked direct evidence until now [11,12]. Other suspected routes of transmission still needed to be clari ed.

Transmission Patterns
The transmission patterns of COVID-19 included community transmission, nosocomial transmission, household transmission, and transmission in closed environments. Transmissions between patients and health workers were in higher proportion during the SARS outbreak, while transmission through medical facilities was higher proportion during the MERS outbreak [13]. In Wuhan, the proportion of severely infected medical workers was higher than the national average [3]. In Italy, 2,629 health workers were reported infected with the COVID-19 before 18th March and accounted for 8.3% of the total number of cases nationwide. The number however increased to 8,358 by 30th March and represented 9% of the country's total number of cases [W11, W12]. In Spain, the number of diagnosed cases among medical workers increased to 5,969 within 6 days and more than 12% of the country's con rmed cases remained among medical workers until March 30th [W12]. Update from another source reported an increase in the number of cases from 12-14% among Spain healthcare workers by 31st March and this was attributed to lack of medical supplies, such as masks and gowns. Other reasons accounting for these high infection rates among medical personnel varied according to different country's circumstances. An Italy study pointed out hospitals as a potential hotspot for infection. Facilities and medical personnel turned into untested vectors and patients [W13, W14]. In the US for example, the reasons that turned hospitals into infection hotspots included the overload of COVID-19 patients and inappropriate management against the pandemic in hospitals [W15]. Similar to the US, 200 medical workers got infected in a county hospital in Romania due to inadequate hospital management. In Egypt, a serious wave of emigration by physicians for years led to patient overload for remaining medical workers and placed them at higher risk of infection through continuous exposure. The emigration wave was purportedly caused by low salary, undesirable working conditions, lack of legal protection, and shortage . Further studies are however required to identify and assess other potential transmission patterns for further prevention, especially since some cases were asymptomatic [15,16]. In addition, patients who were considered cured and no longer needed quarantine still tested RT-PCR positive after 5 to 13 days [17].

Nature history
We systematically pooled data on the incubation period and the reproduction numbers for analysis. The pooled data suggested that the mean incubation period was 5.24 days (95% CI:3.97-6.50, 5 studies), and ranged from 3-7.4 days [W23-W27]. However, the incubation period in some special cases could be as long as 24 days [10]. The pooled results also illustrated that the basic reproduction number (R0) of SARS-CoV-2 was 3.32 (95% CI:3.24-3.39, 14 studies) and varied between 0.6-6.47 [W24-W35, 15]. The results suggested that the transmission ability of SARS-CoV-2 was stronger than SARS (3) and MERS (≤ 1) [W36, W37]. Moreover, the median time from the rst symptom to rst hospital admission was 7 days with the median duration from illness development to severe symptoms development being: 5-8 days for dyspnea, 8-9 days for ARDS, 10.5 days for mechanical ventilation and ICU admission [6,18]. For COVID-19 related deaths, the duration from the onset of symptoms to death averaged 9 days in China 5 and in Italy (median) [W5], and 10 days in South Korea (median) [2].

Mortality
By 21st Apr, 16 nations had reported over 20,000 COVID-19 cases in each of the countries, together contributed to 84.6% of the con rmed cases and 92.4% of death in the world [W38]. The global Case Fatality Rate (CFR) was 7.0% on 21st Apr. However, it was apparently different by country. Among the 16 countries, 6 had over 10% CFR with an apparently increasing trend after reported the over-100 cases (Fig. 2). These countries were France (17.7%), Belgium (14.6%), Italy (13.3%), United Kingdom (13.2%), Netherlands (11.2%), and Spain (10.4%), which were all from the European region. Among the other 10 countries, the CFR of Brazil, Canada, Switzerland, Portugal, and Germany slowly increased to 6.3%, 4.6%, 4.1%, 3.5%, and 3.2%, respectively; the CFR of China, Russia, and Turkey was relatively stable at 5.5%, 0.9%, and 2.4%, respectively. The CFR of Iran and the United States largely uctuated. The CFR of Iran rstly declined to 2.5% on 8th Mar., then increased to 7.9% on 24th Mar., and slowly decreased again and became stable at 6.2% on 21st Apr. The CFR of the United States decreased to 1.1% on 20th Mar. and then slowly increased to 5.4% on 21st Apr. As the pandemic outbreak continued, more surveillance is needed for the CFR of COVID-19.
However, in Italy's report, the most commonly observed symptoms were fever and dyspnea whiles cough, diarrhea and hemoptysis were less common. Overall, 6.4% of patients did not present any symptoms during hospital admissions [W5]. In Europe, one study found olfactory and gustatory disorders were common symptoms, which were totally different from China's ndings [24]. In the US, the gastrointestinal symptoms were shown more frequently than in China [25].
The top 3 common symptoms among mild and severe patients are summarized and displayed in a gure (Fig. 3) [6,10,18,[26][27][28][29][30]. Fever was found to be the most common symptom in all patients. In a study, 43.8% of patients had fever initially and the proportion increased to 87.9% following hospitalization [10]. The body temperatures of 44%-47.1% of patients ranged between 38.1-39.0℃. The higher body temperatures (above 39.0℃) as well as dyspnea and anorexia were more frequent among patients in severe conditions only [6,10,29]. Cough and fatigue on the other hand were more widely reported among mild and severe patients. What's more, another study reported that dyspnea (76%) was the most common symptom among severe patients in the United States [31].

Laboratory ndings and CT Scans
Laboratory ndings Among COVID -19 patients, a decrease in leukocytes such as eosinophil and lymphocyte were commonly reported. This might be because the cytokine storm caused by the novel virus changes the peripheral of white blood cells and immune cells [6, 8-10, 18, 23]. Severe lymphopenia was also common among the dead patients [6,30]. Myocardial zymogram abnormality was found in many patients. For instance, 76% of patients had an increase in lactate dehydrogenase, while 13% of patients had increases in creatine kinase [23]. The level of C-reactive protein was important to evaluate the infection [10]. Most patients were found to have a higher level of C-reactive protein (86%) and serum ferritin (63%) compared to the normal range [23]. The biomarkers related to liver and renal damage were found to be abnormal among COVID-19 patients. The abnormality of liver-related biomarkers was not widespread but yet still common in severe cases [6,9,10,32]. Besides, although only 7% of patients showed renal biomarker abnormalities, renal damage might contribute to the nal multi-organ failure and death outcome [33, 34].
The ICU patients showed higher levels of white blood cells, neutrophil counts, D-dimer, creatine kinase, and creatine with longer prothrombin times [6,10,18]. Compared to patients who survived, the patients who died had higher levels of D-dimer, high-sensitivity cardiac troponin I, serum ferritin, lactate dehydrogenase, IL-6, blood urea, creatinine, white blood cell counts and neutrophil counts. Severe lymphopenia was also common among dead patients [6,30].

Computed Tomography Scan (CT Scan) features
Most patients had GGO and the bilateral lung involvement [6,23,[35][36][37]. One study found that bilateral lung involvement was more frequently shown in the intermediate course and late course, compared to the earlier clinical course [38]. The clinical course could be divided into four stages based on CT scan ndings [37]. In the rst stage (Pre-symptom), GGO, unilateral and multifocal were observed among most patients in this stage [37,38]. In the second stage (symptoms ≤ 1 week), lesions soon developed into bilateral and diffused except for GGO. This stage was considered a period from transition to consolidation. A mixed pattern of transition and consolidation develops during this stage. In the third stage (symptoms 1-2 weeks), the GGO was still common and the consolidation pattern showed. Findings indicated an interstitial change, which was considered as the development of brosis. In the fourth stage (symptom 2-3 weeks), consolidation and mixed patterns were more common, and the GGO started to shrink [37], the consolidation was gradually absorbed among patients who recovered at last [39].
Among ICU patients, the bilateral multiple lobular and subsegmental areas of consolidation were considered typical ndings [6]. Patients in severe condition showed diffuse lesions, with density increasing in both lungs. CT scans showed 'white lung' appearances, indicating the serious in uence the infection has on patients' lung functions [40].

Risk factors
Being old (≥ 65 years old), having comorbidities (e.g. hypertension, diabetes, cardiovascular and cerebrovascular diseases, etc.), and developing complications were three vital risk factors for patients to develop severe conditions [18,26,27]. Findings from multiple studies have shown that patients who are more than 65 years of age, and with comorbidities such as diabetes and heart diseases had a high mortality rate [26,30,[41][42][43]. Late hospitalization and bacterial infections were also considered high risk factors for disease progression [23,27,42]. Smoking history could be a potential risk factor for developing severe conditions [23,27]. People with underlying disorders were considered to be at a high risk of getting infected [W3].

Discussion
Our review identi ed several research gaps. Firstly, large amounts of data for countries outside of China were missing from this review. This is because the rst case of COVID-19 was found in China and lasted for three months, while other countries were still at the early stages of the epidemic, except for South Korea [W38]. Besides, published studies of coronavirus in China grew rapidly since the SARS outbreak in 2002, maintaining a top 2 position in the world [44]. Secondly, although mother-to-child transmission was conceded to be a possible transmission route [W39], no direct evidence has proved this and further investigation is needed [3,45]

Conclusions
By collecting and analyzing published COVID-19 articles, we found the average incubation period of COVID-19 was 5.24 days, the R0 was 3.32, indicated a stronger transmission ability than SARS and MERS. The common symptoms had a little difference in Europe and Asia. Body temperatures above 39.0 ℃, dyspnea, and anorexia were more common symptoms in severe patients. Age over 60 years old, having co-morbidities, and developing complications were high-risk factors developed into severe conditions or death. A decrease in leukocytes was common, the abnormal lab biomarkers of cardiac, liver, and renal were found. Observed from the CT ndings, GGO and the bilateral lung involvement were common, the longer the time after symptoms onset, the more consolidated lung lesions were found.

Availability of data and materials
The key information and data generated and/or analyzed during this study were included in this article and/or its supplementary information les.

Competing Interests
Dr. Weiming Tang is the Associate Editor of this journal. Authors' contributions WT designed the study protocol. YX, ZW, and HL did the literature search. The titles, abstracts, and full texts were screened and selected by YX, ZW, and HL. The data were extracted and analysed by YW, ZW and HL. YX, ZW and HL drafted the manuscript. YX, ZW, GM, DW, and WT edited the draft. All authors read and approved the nal manuscript. Case fatality rate of countries reported over 20,000 cases, 2020* *Data was collected until 21 April 2020.
The CFR of a country was not included on those dates when the country reported less than 100 cases, with the consideration that the CFR may not be reliable if the size of infected population was small