Discussion
Up to now, this is the first meta-analysis to explore severe Covid-19 associated clinical, laboratory and imaging factor when compared with non-severe Covid-19. By systemically and comprehensively reviewed the current evidence published, 14 studies with a total of 2,566 individuals (771 in Severe group and 1,795 in Non-severe group) were eligible for this meta-analysis[14-27], which retrieved the largest sample size when compared with studies on the same topic. Overlapping patient was checked by examining the first author of article and the origin of patients, since we recognized that different articles might report of the same patients.
Currently, the National Health Commission (NHC) issued the China Guidelines for the Diagnosis and Treatment Plan of Novel Coronavirus (COVID-19), which defined the degree of severity of Covid-19 (i.e., mild, common, severe and critical). As we all know, treatment algorithm of Covid-19 depended on illness severity. Most severe and critical patients required oxygen therapy and a minority of the patients needed invasive ventilation or even extracorporeal membrane oxygenation. Moreover, there were some patient who developed worsening respiratory failure and acute respiratory distress syndrome (ARDS) rapidly that required intubation[36]. According to epidemiological investigation, severe illness occurred in 15.7% of the Covid-19 patients after admission to a hospital. As the clinical spectrum of COVID-19 ranges widely from mild illness to ARDS with a high risk of mortality, there is an urgent need for research to identify early markers of disease severity, which is of great value for clinician to diagnosis of the severity of Covid-19 rapidly and exactly.
Though statistical analysis, it was demonstrated that patients in Severe Covid-19 group were older and had a greater number of comorbid conditions (e.g., hypertension, diabetes and heart disease) than Non-severe group. Compromised respiratory status on admission (e.g., COPD) was also associated with severe illness. This suggests that age and comorbidity may be risk factors for poor outcome. Meanwhile, severe 2019-nCoV infection is more likely to affect males. These data was consistent with the recent report that showed 2019-nCoV infection is more likely to affect males [37]. What’s more, our outcome did not support that smoking was associated with severity of COVID-19 illness. Consistently, Lippi et al conducted a meta-analysis of current evidence and concluded that active smoking does not apparently seem to be significantly associated with enhanced risk of progressing towards severe disease in Covid-19, which further confirmed our outcome[38].
Common symptoms of Covid-19 at onset of illness were fever, dry cough, expectoration, myalgia, fatigue, and dyspnea [1]. However, some patients presented initially with atypical symptoms, such as diarrhea and nausea [39, 40]. By statistical synthesis the data on common sign and symptoms, the incidence of fever, expectoration, headache, fatigue, myalgia and dyspnea were more common in Severe group than in Non-severe group. However, only the incidence of dyspnea was statistically different across groups. Thus, patient presented with dyspnea should gain more caution for which might be severe Covid-19. And this outcome was consistent with outcome found in ICU patient[41].
Accumulating evidence suggests that a subgroup of patients with severe COVID-19 might have a cytokine storm syndrome[42]. In our study, compared with non-severe patients, severe Covid-19 patients had numerous laboratory abnormalities. By meta-analysis of current evidence, depressed total lymphocytes were observed in this article. These abnormalities suggest that 2019-nCoV infection may be associated with cellular immune deficiency. And these laboratory abnormalities are similar to those previously observed in patients with MERS-CoV and SARS-CoV infection[43].
Individuals with severe Covid-19 might present with bilateral (95.8%) or unilateral (30.5%) lung pathological changes, ground-glass opacities (100%), consolidation (76.9%) and bronchial wall thickening (56.4%) in chest CT. However, no statistical difference was revealed when compared with Non-severe group. Although reticulation (30.8%), intrathoracic lymph node enlargement (20.5%) and pleural effusions (30.8%) were relatively rarely seen, meta-analysis revealed that patients with reticulation, intrathoracic lymph node enlargement and pleural effusions in chest CT were associated with more likelihood to be a severe Covid-19. The outcome was further confirmed in a study carried by Yuan eta al which investigated the association of radiologic findings with mortality of patients infected with Covid-19[16].
The results should be viewed with recognition of limitations inherent in this study. Firstly, although a broad review scope provides us with a larger sample size and finally adequate statistical power to detect a risk factor, three articles reporting data comparing clinical characteristic between severe Covid-19 and non-severe Covid-19 was excluded for overlapping patients [3, 35, 44]. One article that did not reported the origin of patient was also excluded from this meta-analysis [29], which resulted in relative small sample size. However, their outcome further confirmed our conclusion.
Secondly, all eligible studies came from China, since first Covid-19 was identified in Wuhan, China. Data in other country was not acceptable right now. Thus, the outcome of our study could not be considered conclusive on this topic. An update of this article is necessary when needed.
Thirdly, more and more articles on Covid-19 were published every day. There might be lots of article evaluating the clinical difference across severe and non-severe Covid-19 unpublished. And funnel plot of this meta-analysis revealed that publication bias might exist. Thus, it is necessary for clinicians to interpret our outcome carefully.
In conclusion, it was demonstrated that older males manifested with dyspnea whose blood routine test revealed lymphopenia should gain more caution for which might be severe Covid-19. Patients with comorbidity, such as hypertension, diabetes and heart disease were more susceptible to severe Covid-19. Compromised respiratory status on admission (e.g., COPD) was also associated with severe illness. Specially, although reticulation, intrathoracic lymph node enlargement and pleural effusions were relatively rarely seen, meta-analysis revealed that patients with such presentations in chest CT were associated with more likelihood to be a severe Covid-19.
Although lots of risk factors were filtrated in this article, exploration of predicted value of these factors in severe Covid-19 patients was impossible with aggregated data extracted from published studies. Further diagnostic article evaluating how to differentiate severe from non-severe Covid-19 with the manifestation in chest CT and study evaluating the relation across clinical characteristic and severity of Covid-19 with the help of logistic regression analysis is needed.