The Novel Coronavirus Disease (COVID-19): A PRISMA Systematic Review and Meta-Analysis of Clinical and Paraclinical Characteristics
An outbreak of pneumonia, caused by a novel coronavirus (SARS-CoV-2), was identified in China in December 2019. This virus expanded worldwide, causing global concern. Although clinical, laboratory, and imaging features of COVID-19 are characterized in some observational studies, we undertook a systematic review and meta-analysis to assess the frequency of these features. We did a systematic review and meta-analysis using three databases to identify clinical, laboratory, and computerized tomography (CT) scanning features of rRT-PCR confirmed cases of COVID-19. Data for 3420 patients from 30 observational studies were included. Overall, the results showed that fever (84.2%, 95% CI 82.6-85.7), cough (62%, 95% CI 60-64), and fatigue (39.4%, 95% CI 37.2-41.6%) are the most prevalent symptoms in COVID-19 patients. Increased CRP level, decreased lymphocyte count, and increased D-dimer level were the most common laboratory findings. Among COVID-19 patients, 92% had a positive CT finding, most prevalently ground-glass opacification (GGO) (60%, 95% CI 58-62) and peripheral distribution opacification (64%, 95% CI 60-69). These results demonstrate the clinical, paraclinical, and imaging features of COVID-19.
In December 2019, the first case of unknown origin pneumonia was identified in Wuhan, the capital city of Hubei Province. By January 7, 2020, Chinese scientists had isolated a novel virus belongs to coronaviruses family and classified it as a type of RNA virus .
Although the outbreak has been started from a primary zoonotic virus transmission in a large seafood market (Huanan Seafood Market), person-to-person transmission of the virus started a pandemic involving 197 countries [2, 3].
The clinical outcomes of SARS-CoV-2 infection are various, including asymptomatic infection, mild upper respiratory tract illness, severe viral pneumonia, and even death. Patients are presented with various clinical manifestations such as fever, dyspnea, and cough .
Initially, some studies have observed particular imaging patterns on chest radiography and computed tomography in COVID-19 patients . As our knowledge increased, recent studies claimed that the sensitivity of CT scan is higher compared to rRT-PCR in the diagnosis of COVID-19 .
Laboratory findings are essential in order to evaluate patients’ complications and triaging them . Complete blood count as an easy and affordable test detects disorders such as leukopenia, anemia, and thrombocytopenia that are contributed to patients’ prognosis . In response to inflammation induced by COVID-19, acute-phase reactants may increase or decrease . These factors may contribute to the patients’ outcomes. This meta-analysis is aimed at measuring the most common clinical, laboratory, and imaging findings among COVID-19 patients.
2.2. Eligibility Criteria
In this study, all included patients were confirmed using real-time reverse transcriptase-polymerase chain reaction (rRT-PCR). All searched articles were cross-sectional studies, reporting descriptive data, and no language restrictions were conducted. All articles published before drafting the manuscript have been included. Review articles, opinion articles, and letters not presenting original data were excluded from the analysis.
2.3. Information Sources and Search Strategy
Three systematic searches were performed using Medline/PubMed, Scopus, and Web of Science. Systematic search was conducted prior to March 25, 2020, and three independent researchers evaluated all papers. The search was conducted based on the following keywords, and all studies were divided in three groups: (1) For clinical characteristics group: (“clinical manifestation” AND COVID-19) or (“clinical manifestation” AND 2019-nCoV) or (“clinical manifestation” AND COVID) or (“clinical manifestation” AND Corona) or (“clinical characteristics” AND COVID-19) or (“clinical characteristics” AND 2019-nCoV) or (“clinical characteristics” AND COVID) or (“clinical characteristics” AND corona). (2) For laboratory findings group: (liver AND COVID-19) or (liver AND 2019-nCoV) or (liver AND COVID) or (liver AND Corona) or (“blood gas” AND COVID-19) or (“blood gas” AND 2019-nCoV) or (“blood gas” AND COVID) or (“blood gas” AND corona). (3) For imaging studies group: (COVID-19 AND radiography) or (2019-nCoV AND radiography) or (Corona AND radiography) or (COVID AND radiography) or (COVID-19 AND radiographic) or (2019-nCoV AND radiographic) or (Corona AND radiographic) or (COVID AND radiographic) or (COVID-19 AND CT) or (2019-nCoV AND CT) or (Corona AND CT) or (COVID AND CT) or (COVID-19 AND “computed tomography”) or (2019-nCoV AND “computed tomography”) or (Corona AND “computed tomography”) or (COVID AND “computed tomography”) or (CBC AND corona) or (CBC AND COVID) or (CBC AND COVID-19) or (CBC AND 2019-nCoV).
2.4. Study Selection
In the initial search, we assessed the title and abstract, followed by a full-text evaluation based on previously described inclusion and exclusion criteria. When two articles reported one patient’s characteristics, we merged all reported data and assumed as a single individual. Descriptive studies reporting clinical symptoms, laboratory, and radiological findings were used to perform a meta-analysis. The characteristics of the included studies are shown in Table 1. The modified appraisal tool for cross-sectional studies (AXIS) was used to determine the methodological quality of the research designs of the included studies (Table 2). AXIS is used to assess research papers systematically and to judge the reliability of the study being presented in the paper. It also helps in assessing the worth and relevance of the study. Studies with total scores of ten or less were excluded.
2.5. Data Collection Process and Data Items
Three independent researchers filled data extraction forms containing study type, journal, publication date, sample size, age, gender, clinical characteristics, laboratory, and radiological findings. Conflicts were resolved by another researcher.
2.6. Assessment of Methodological Quality and Risk of Bias
Publication bias was assessed with a funnel plot for the standard error and considering that the interpretation of the plot is subjective (Figures 4–6). Also, bias was quantified by using the Egger regression test.
Sensitivity analysis and adjusting for risk bias were performed by the attractive test (trim and fill method). We initially identified and trimmed the asymmetric (missing) studies, followed by estimating the unbiased summary effect. Sensitivity analysis was also performed using the “Remove-One” analysis by running the analysis with each of the studies removed. The result of the impact of each study on the pooled estimate is shown in the forest plot (Figure 7).
The percentage of total variation across studies (heterogeneity) was measured by the inconsistency index tool ( squared). The squared index measured for each of the clinical characteristics, imaging studies, and laboratory findings groups. squared index value in the ranges of <25%, 25–50%, 50–75%, and >75% was interpreted as low, moderate, high, and very high heterogeneity, respectively .
We conducted a random-effects analysis because it was assumed that some of the included studies did not share a common effect size (heterogeneity). Findings in each group are summarized as forest plots in Figures 8–10.
2.7. Statistical Approach
Effect size pooled estimate for imaging and clinical data was measured based on event rate, logit event rate, and standard error.
Considering that the computational index of laboratory data was median in order to meta-analyze them in CMA v.2. software, we use the following formula:
Estimating the mean and variance from the median: where is the median, is the smallest value (minimum), is the largest value (maximum), and is the size of the sample .
The meta-analysis was performed using STATA, the software OpenMeta[Analyst], and Comprehensive Meta-Analysis Software (CMA) ve.2. Pooled estimate and their 95% confidence intervals (95% CIs) were used to summarize the weighted effect size for each study grouping variable.
3.1. Study Selection and Characteristics
Two hundred seven articles were included based on a search strategy, which are previously described (Table 1). The full text of 65 articles was evaluated after the title and abstract assessment. Twenty-four articles were excluded due to inadequate data. Finally, the meta-analysis was performed on 30 articles (three different subjects). The article’s data summary is reported in Table 2. Also, demographic characteristics and comorbidities of patients participated in the included studies are demonstrated in Table 3.
In this study, we evaluate 30 articles. All papers were from China, and 3420 individual’s data were evaluated. All studies were cross-sectional, and 27 variables were included.
Evaluating the heterogeneity of the studies indicated that in the clinical characteristics and imaging studies groups, the combined effect of the squared index is high (68.43, 68.53). While the laboratory findings groups combined effect of the squared index is considered low (6.12). squared index for each of the outcomes is shown in Tables 4–7.
3.3. Publication Bias and Sensitivity Analysis
The Funnel plot for clinical characteristics group studies is almost symmetric confirmed by the Egger regression test (, value = 0.20). By using the random-effects model, the summary estimate and 95% confidence interval for the combined studies is 0.25 (0.22, 0.29). These findings indicated no publication bias in the clinical characteristics group (Figure 4).
The funnel plots in the imaging studies group and findings studies seem asymmetric and skewed (Figures 5 and 6). Also, the Egger regression test has indicated an intercept of 1.55, and value of 0.01 for the laboratory findings group and an intercept of 0.94 and value of 0.04 for the imaging studies group.
In the laboratory findings group, using the random-effects model, the point estimate and 95% confidence interval for the combined studies is 3.01 (2.22, 3.80). Using trim and fill (four trimmed studies), the imputed summary estimate is 3.30 (2.30, 3.38) (Figure 5).
In the imaging studies group, the summary estimate and 95% confidence interval for the combined studies is 0.51 (0.48, 0.54). Using trim and fill, the imputed (four trimmed studies) summary estimate is 0.50 (0.47, 0.53) (Figure 6).
In conclusion, the finding of trim and fill analysis has indicated only minimal changes, which do not seem to be a threat to the validity of the effect size estimates.
Sensitivity analysis by using the “Remove-One” analysis did not show any change in the combined effect of the clinical characteristics and imaging studies groups after removing any of the studies. However, in the laboratory findings group, removing only one of the studies (Fan et al.) changed the combined effect significantly (Figure 7). The combined effect before and after removing Fan et al. study is presented in Tables 5 and 6.
3.4. Clinical Characteristics Group
According to clinical manifestations, fever (84.3%, 95% CI 78.6-88.7), cough (60.1%, 95% CI 53.5-66.4), and fatigue (39.4%, 95% CI 29.1-50.8) are the most prevalent clinical symptoms among patients (Table 4) (Figure 8).
3.5. Laboratory Findings Group
Laboratory studies show increased level in following tests: CRP (10.78 mg/l, 95% CI 6.44-15.11) with the normal range of 0-3.0 mg/l, D-dimer (567.89 ng/ml, 95% CI 348.15-787.62) with the normal range of 0-500 ng/ml, LDH (258.56 U/l, 95% CI 206.84-310.29) with the normal range of 135-250 U/l, and procalcitonin (0.17 ng/ml, 95% CI 0.01-0.32) which is normally less than 0.05 ng/ml in a healthy individual.
Also, the level of some laboratory factors is lower than normal, such as lymphocyte (, 95% CI 0.73-1.26) and albumin (38.61 g/l, 95% CI 35.75-41.48) concerning the normal range of and 40-55 g/l, respectively (Tables 5 and 6) (Figure 9).
3.6. Imaging Studies Group
Among all patients infected by SARS-CoV-2 (confirmed by RT-PCR), 85% had abnormalities in CT scans. In most of them, the bilateral pneumonia was dominant (79.4%, 95% CI 66.9-88.1). Ground-glass opacification (GGO) (69.8%, 95% CI 60.3-77.9), peripheral distribution (66.8%, 95% CI 50.0-80.2), and consolidation (37.8%, 95% CI 26.4-50.8) in those with CT scan results are presented (Table 7) (Figure 10).
From December 2019, more than 500,000 cases of new unknown origin pneumonia have been confirmed all over the world . It was primarily known as 2019-nCov; then, WHO decided to name this novel coronavirus “SARS-CoV-2” . COVID-19 is a severe condition that is compromising the health condition of people in all countries worldwide . Identifying the various characteristics of this infection is vital for controlling the outbreak in different countries . Clinical, laboratory, and imaging findings are essential to evaluate the different aspects of infection . Different outcomes of COVID-19 (from an asymptomatic infection to death) and contagiousness of this virus, even in its incubation period , emphasize why discovering different characteristics are crucial in controlling this pandemic.
In this systematic review and meta-analysis, we describe the most common clinical data on COVID-19 confirmed cases that were published during the first months of the outbreak. We analyzed 2422, rRT-PCR confirmed patients, for different clinical manifestations. Our findings are robust due to the pooled results after combining all the studies’ data.
As expected from initial studies in China, COVID-19 patients presented predominantly with cough and fever, as well as headache, diarrhea, and fatigue, among other clinical features . This was consistently found in many of the included studies [19, 20]. Fever frequency is similar in other β-CoV-associated infections such as SARS and MERS, but studies showed that the cough frequency is higher in SARS and COVID-19 than MERS (<50%) [21, 22]. In SARS and MERS, diarrhea is reported in about a quarter of patients, but our data shows that only 6 percent of COVID-19 patients present with diarrhea (Table 4). Our data also suggest that about 11 percent of patients are presented with headache as a symptom. Unlike SARS, which is well characterized in the two-stage clinical course of the disease , COVID-19 still needs further definition to identify the disease process.
Studies on epidemiological features of COVID-19 showed that about 80 percent of patients are asymptomatic or are presented with mild manifestations [24, 25], but almost all of the patients included in our study had moderate-to-severe characteristics. It seems that fever and cough are the most common clinical features among moderate-to-severe patients (Table 4).
Studies show different laboratory abnormalities in COVID-19 patients, such as hypoalbuminemia or elevated inflammatory markers . However, our data suggest that C-reactive protein is the most elevated factor among infected cases (Tables 5 and 6). D-dimer, LDH, and procalcitonin are also elevated in patients, which confirmed that measuring inflammatory markers are essential to investigate new cases . Also, seven studies showed lymphopenia and albuminuria as other common laboratory findings. Data from new studies suggest lymphocytopenia or an increase in WBC as prognostic factors in COVID-19 patients. Studies on the SARS outbreak in 2003 indicate that lymphopenia, leukopenia, and thrombocytopenia, elevated levels of LDH, alanine transaminase (ALT), AST, and creatine-kinase are the most affected laboratory findings .
Nevertheless, not significantly seen in COVID-19, the novel corona virus can affect the liver and other organs . AST and ALT are normal in most cases, but impaired liver function tests are associated with poor prognosis and higher mortality rates [30, 31]. Coagulation function tests (such as INR) are affected in the prognosis of this infection . Lymphopenia in COVID-19 patients suggests that this virus might act on lymphocytes (mainly T cells), but some studies suggest that B cells are also affected [26, 33].
CT scan is one of the most useful methods to diagnose the respiratory tract diseases diagnosis. CT scan high sensitivity and availability makes it one of the most common tests for lung disease screening [34, 35]. In COVID-19 patients, different results could present in the early stages of infection [36, 37]; even some studies demonstrate that CT scan sensitivity is higher than rRT-PCR . Our data shows that 92% of rRT-PCR confirmed cases had abnormal CT scan results, which suggest CT scan as a reliable method. As seen in Table 7, CT scan meta-analysis outcomes are performed in random-effects analyses. Our meta-analysis on 15 studies showed that ground-glass opacification (GGO) and peripheral distribution are seen on 69.8% and 66.8% of patients, respectively. 79.4% of the patients had bilateral involvements, which is contributed to poor prognosis. CT scan is useful in monitoring the treatment, and it is crucial in classifying patients and identifying who should be treated with aggressive treatments . Other findings such as consolidation or reverse halo or atoll sign are reported in some studies , which were not included in our analysis.
This review has several limitations. Few studies are available on COVID-19, and most of them are from China. Many countries such as Italy, the United States, and Iran reported several new COVID-19 patients, but data about clinical characteristics or laboratory findings are limited. By publishing more studies worldwide, researchers are going to get a more comprehensive understanding of COVID-19. Patients’ detailed information, especially in clinical outcomes, was unavailable in most studies at the time of analysis. In this study, we used random-effects model for analysis in all three groups. In comparison to the fixed model, random model findings have wider confidence intervals and less accurate results, although heterogeneity in included studies could not be considered in the fixed model.
COVID-19 presents in the majority of cases with fever and cough. Laboratory findings such as elevated inflammatory markers can assist the diagnosis. Other laboratory indices, such as AST, ALT, or INR, are also affected in these patients. 92% of the RT-PCR confirmed that patients have abnormalities in CT scan most frequently bilateral involvement. Additional research with higher sample sizes is needed in order to describe the patients’ characteristics more precisely.
|COVID-19:||Corona virus disease of 2019|
|rRT-PCR:||Real-time reverse transcription-polymerase chain reaction|
|CT scan:||Computed tomography scan|
|SARS-CoV-2:||Severe acute respiratory syndrome corona virus #2|
|PRISMA:||Protocol based on the transparent reporting of systematic reviews and meta-analysis|
|CBC:||Complete blood count|
|MERS:||Middle East respiratory syndrome|
|SARS:||Severe acute respiratory syndrome|
|WBC:||White blood cell|
|INR:||International normalized ratio.|
Conflicts of Interest
The authors declare that they have no competing interest regarding the publication of this paper.
HH managed the group and participated in preparing the manuscript. SS participated in meta-analysis. SM and NT participated in the systematic search and substantially drafted the manuscript. PM revised the manuscript critically. All authors read and approved the final version of the manuscript.
World Health Organization, “Laboratory testing for coronavirus disease 2019 (COVID-19) in suspected human cases: interim guidance, 2 March 2020,” 2020.View at: Google Scholar
Y. X. Li, W. Wu, T. Yang et al., “Characteristics of peripheral blood leukocyte differential counts in patients with COVID-19,” Zhonghua Nei Ke Za Zhi, vol. 59, article E003, 2020.View at: Google Scholar
X. Luo, W. Zhou, X. Yan et al., “Prognostic value of C-reactive protein in patients with COVID-19,” 2020, https://www.medrxiv.org/content/10.1101/2020.03.21.20040360v1.View at: Google Scholar
WHO, “Coronavirus disease 2019 (COVID-19) situation report – 69,” Tech. Rep., 2020, https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200329-sitrep-69-covid-19.pdf?sfvrsn=8d6620fa_2.View at: Google Scholar
A. E. Gorbalenya, S. C. Baker, R. S. Baric et al., “Severe acute respiratory syndrome-related coronavirus–the species and its viruses, a statement of the Coronavirus Study Group,” 2020, https://www.biorxiv.org/content/10.1101/2020.02.07.937862v1.View at: Google Scholar
J. Wu, J. Liu, X. Zhao et al., “Clinical characteristics of imported cases of COVID-19 in Jiangsu Province: a multicenter descriptive study,” Clinical Infectious Diseases, vol. 10, 2020.View at: Google Scholar
S. Meo, A. M. Alhowikan, T. al-Khlaiwi et al., “Novel coronavirus 2019-nCoV: prevalence, biological and clinical characteristics comparison with SARS-CoV and MERS-CoV,” European Review for Medical and Pharmacological Sciences, vol. 24, no. 4, pp. 2012–2019, 2020.View at: Publisher Site | Google Scholar
H. Yu, J. Shao, Y. Guo et al., “Data-driven discovery of clinical routes for severity detection in COVID-19 pediatric cases,” 2020, https://www.medrxiv.org/content/medrxiv/early/2020/03/10/2020.03.09.20032219.full.pdf.View at: Google Scholar
L. Tan, Q. Wang, D. Zhang et al., “Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study,” 2020, https://www.medrxiv.org/content/10.1101/2020.03.01.20029074v1.View at: Google Scholar
J. Gong, H. Dong, S. Q. Xia et al., “Correlation analysis between disease severity and inflammation-related parameters in patients with COVID-19 pneumonia,” 2020, https://www.medrxiv.org/content/10.1101/2020.02.25.20025643v1.View at: Google Scholar
N. Yao, S. N. Wang, J. Q. Lian et al., “Clinical characteristics and influencing factors of patients with novel coronavirus pneumonia combined with liver injury in Shaanxi region,” Zhonghua gan zang bing za zhi= Zhonghua ganzangbing zazhi= Chinese Journal of Hepatology, vol. 28, 2020.View at: Google Scholar
J. Wang, J. Liu, Y. Wang et al., “Dynamic changes of chest CT imaging in patients with corona virus disease-19 (COVID-19),” Journal of Zhejiang University (Medical Science), vol. 49, no. 1, 2020.View at: Google Scholar
G.-Q. Qian, N.-B. Yang, F. Ding et al., “Epidemiologic and clinical characteristics of 91 hospitalized patients with COVID-19 in Zhejiang, China: a retrospective, multi-centre case series,” 2020, https://www.medrxiv.org/content/10.1101/2020.02.23.20026856v2.View at: Google Scholar
Y. Liu, W. Sun, L. Chen, Y. Wang, L. Zhang, and L. Yu, “Clinical characteristics and progression of 2019 novel coronavirus-infected patients concurrent acute respiratory distress syndrome,” 2020, https://www.medrxiv.org/content/10.1101/2020.02.17.20024166v1.article-info?versioned=true.View at: Google Scholar
G. Chen, W. G. Di Wu, Y. Cao et al., “Clinical and immunologic features in severe and moderate forms of coronavirus disease 2019,” 2020, https://www.medrxiv.org/content/10.1101/2020.02.16.20023903v1.View at: Google Scholar