Clinical presentations, laboratory and radiological findings, and treatments for 11,028 COVID-19 patients: a systematic review and meta-analysis

This systematic review and meta-analysis investigated the comorbidities, symptoms, clinical characteristics and treatment of COVID-19 patients. Epidemiological studies published in 2020 (from January–March) on the clinical presentation, laboratory findings and treatments of COVID-19 patients were identified from PubMed/MEDLINE and Embase databases. Studies published in English by 27th March, 2020 with original data were included. Primary outcomes included comorbidities of COVID-19 patients, their symptoms presented on hospital admission, laboratory results, radiological outcomes, and pharmacological and in-patient treatments. 76 studies were included in this meta-analysis, accounting for a total of 11,028 COVID-19 patients in multiple countries. A random-effects model was used to aggregate estimates across eligible studies and produce meta-analytic estimates. The most common comorbidities were hypertension (18.1%, 95% CI 15.4–20.8%). The most frequently identified symptoms were fever (72.4%, 95% CI 67.2–77.7%) and cough (55.5%, 95% CI 50.7–60.3%). For pharmacological treatment, 63.9% (95% CI 52.5–75.3%), 62.4% (95% CI 47.9–76.8%) and 29.7% (95% CI 21.8–37.6%) of patients were given antibiotics, antiviral, and corticosteroid, respectively. Notably, 62.6% (95% CI 39.9–85.4%) and 20.2% (95% CI 14.6–25.9%) of in-patients received oxygen therapy and non-invasive mechanical ventilation, respectively. This meta-analysis informed healthcare providers about the timely status of characteristics and treatments of COVID-19 patients across different countries. PROSPERO Registration Number: CRD42020176589

Data analysis. Three authors (CW, EHMT and CHA) extracted data using a standardized spreadsheet to record the article type, country of origin, surname of first author, year of publications, sample size, demographics, comorbidities, symptoms, laboratory and radiology results, pharmacological and non-pharmacological treatments.
We aggregated estimates across 90 eligible studies to produce meta-analytic estimates using a random-effects model. For dichotomous outcomes, we estimated the proportion and its respective 95% confidence interval. For laboratory parameters as continuous outcomes, we estimated the mean and standard deviation from the median and interquartile range if the mean and standard deviation were not available from the study 14 , and calculated the mean and its respective 95% confidence intervals. Random-effect models on DerSimonian and Laird method were adopted due to the significant heterogeneity, checked by the I 2 statistics and the p values. I 2 statistic of < 25%, 25-75% and ≥ 75% is considered as low, moderate, high likelihood of heterogeneity. Pooled estimates were calculated and presented by using forest plots. Publication bias was estimated by Egger's regression test. Funnel plots of outcomes were also presented to assess publication bias.
All statistical analyses were conducted using the STATA Version 13.0 (Statacorp, College Station, TX). The random effects model was generated by the Stata packages 'Metaprop' for proportions 15 and 'Metan' for continuous variables 16 .

Results
The selection and screen process are presented in Fig. 1. A total of 241 studies were found by our searching strategy (71 in PubMed and 170 in Embase). 46 records were excluded due to duplication. After screening the abstracts and titles, 100 English studies were with original data and included in full-text screening. By further excluding 10 studies with not reporting symptoms presentation, laboratory findings, treatment and therapeutics, 90 studies 17-106 and 76 studies with more than one COVID-19 case 17 www.nature.com/scientificreports/ in China. Newcastle-Ottawa Quality Assessment Scale has been used to assess study quality of each included cohort study 107 . 30% (27/90) of included studies had satisfactory or good quality. The summary of the included study is shown in Table 1.
The prevalence of radiological outcomes and non-pharmacological treatments were presented in Fig. 6. Radiology findings detected chest X-ray abnormalities, with 74

Discussion
This meta-analysis reveals the condition of global medical community responding to COVID-19 in the early phase. During the past 4 months, a new major epidemic focus of COVID-19, some without traceable origin, has been identified. Following its first identification in Wuhan, China, the virus has been rapidly spreading to Europe, North America, Asia, and the Middle East, in addition to African and Latin American countries. Three months since Wuhan CDC admitted that there was a cluster of unknown pneumonia cases related to Huanan Seafood Market and a new coronavirus was identified as the cause of the pneumonia 108 , as on 1 April, 2020, there have been 858,371 persons confirmed infected with COVID-19, affecting 202 countries and territories around the world. Although this rapid review is limited by the domination of reports from patients in China, and the patient population is of relative male dominance reflecting the gender imbalance of the Chinese population 109 , it provides essential information.
In this review, the pooled mean age was 45.8 years. Similar to the MERS-CoV pandemic 110 , middle-aged adults were the at-risk group for COVID-19 infections in the initial phase, which was different from the H1N1 influenza pandemic where children and adolescents were more frequently affected 111 . Biological differences may affect the clinical presentations of infections; however, in this review, studies examining the asymptomatic COVID-19 infections or reporting any previous infections were not included. It is suggested that another systematic review should be conducted to compare the age-specific incidence rates between the pre-pandemic and post-pandemic periods, so as to understand the pattern and spread of the disease, and tailor specific strategies in infection control.
Both sexes exhibited clinical presentations similar in symptomatology and frequency to those noted in other severe acute respiratory infections, namely influenza A H1N1 112 and SARS 113,114 . These generally included fever, new onset or exacerbation of cough, breathing difficulty, sore throat and muscle pain. Among critically ill patients usually presented with dyspnoea and chest tightness 22,29,39,72 , 141 (4.6%) of them with persistent or progressive hypoxia resulted in the requirement of intubation and mechanical ventilation 115 , while 194 (6.4%) of them required non-invasive ventilation, yielding a total of 11% of patients requiring ventilatory support, which was similar to SARS 116 .
The major comorbidities identified in this review included hypertension, cardiovascular diseases and diabetes mellitus. Meanwhile, the percentages of patients with chronic renal diseases and cancer were relatively low. These chronic conditions influencing the severity of COVID-19 had also been noted to have similar effects in other respiratory illnesses such as SARS, MERS-CoV and influenza 117,118 . Higher mortality had been observed among older patients and those with comorbidities.
Early diagnosis of COVID-19 was based on recognition of epidemiological linkages; the presence of typical clinical, laboratory, and radiographic features; and the exclusion of other respiratory pathogens. The case definition had initially been narrow, but was gradually broadened to allow for the detection of more cases, as milder cases and those without epidemiological links to Wuhan or other known cases had been identified 119,120 . Laboratory investigations among COVID-19 patients did not reveal specific characteristics-lymphopenia and elevated inflammatory markers such as CRP are some of the most common haematological and biochemical abnormalities, which had also been noticed in SARS 121 . None of these features were specific to COVID-19. Therefore, diagnosis should be confirmed by SARS-CoV-2 specific microbiological and serological studies, although initial management will continue to be based on a clinical and epidemiological assessment of the likelihood of a COVID-19 infection.
Radiology imaging often plays an important role in evaluating patients with acute respiratory distress; however, in this review, radiological findings of SARS-CoV-2 pneumonia were non-specific. Despite chest radiograph www.nature.com/scientificreports/ www.nature.com/scientificreports/ usually revealed bilateral involvement and Computed Tomography usually showed bilateral multiple ground-glass opacities or consolidation, there were also patients with normal chest radiograph, implying that chest radiograph might not have high specificity to rule out pneumonia in COVID-19.
Limited clinical data were available for asymptomatic COVID-19 infected persons. Nevertheless, asymptomatic infection could be unknowingly contagious 122 . From some of the official figures, 6.4% of 150 non-travelrelated COVID-19 infections in Singapore 123 , 39.9% of cases from the Diamond Princess cruise ship in Japan 124 , and up to 78% of cases in China as extracted on April 1st, 2020, were found to be asymptomatic 122 . 76% (68/90) studies based on hospital setting which provided care and disease management to symptomatic patients had limited number of asymptomatic cases of COVID-19 infection. This review calls for further studies about clinical data of asymptomatic cases. Asymptomatic infection intensifies the challenges of isolation measures. More global reports are crucially needed to give a better picture of the spectrum of presentations among all COVID-19 infected persons. Also, public health policies including social and physical distancing, monitoring and surveillance, as well as contact tracing, are necessary to reduce the spread of COVID-19.
Concerning potential treatment regime, 62.4% of patients received antivirals or interferons (including oseltamivir, lopinavir-ritonavir, interferon alfa), while 63.9% received antibiotics (such as moxifloxacin, and ceftriaxone). In this review, around one-third of patients were given steroid, suggestive as an adjunct to IFN, or sepsis management. Interferon and antiviral agents such as ribavirin, and lopinavir-ritonavir were used during SARS, and the initial uncontrolled reports then noted resolution of fever and improvement in oxygenation and radiographic appearance 113,125,126 , without further evidence on its effectiveness. At the time of manuscript preparation, there has been no clear evidence guiding the use of antivirals 127 . Further research is needed to inform clinicians of the appropriate use of antivirals for specific groups of infected patients.
Limitations of this meta-analysis should be considered. First, a high statistical heterogeneity was found, which could be related to the highly varied sample sizes (9 to 4226 patients) and study designs. Second, variations of follow-up period may miss the event leading to heterogeneity. In fact, some patients were still hospitalized in the included studies. Third, since only a few studies had compared the comorbidities of severe and non-severe patients, sensitivity analysis and subgroup analysis were not conducted. Fourthly, the frequency and severity of signs and symptoms reported in included studies, primarily based on hospitalized COVID-19 patients were over-estimated. Moreover, different cutoffs for abnormal laboratory findings were applied across countries, and counties within the same countries. Lastly, this meta-analysis reviewed only a limited number of reports written in English, with a predominant patient population from China. This review is expected to inform clinicians of the epidemiology of COVID-19 at this early stage. A recent report estimated the number of confirmed cases in     www.nature.com/scientificreports/ www.nature.com/scientificreports/ China could reach as high as 232,000 (95% CI 161,000, 359,000) with the case definition adopted in 5th Edition. In this connection, further evidence on the epidemiology is in imminent need. www.nature.com/scientificreports/