This is the first meta-analysis that avoids the phenomenon of included cases duplication, which compares severe and non-severe COVID-19 in the field of demographic features, clinical symptoms comorbidities, complications and outcomes. Based on 4881 laboratory-confirmed cases with COVID-19 in mainland China from 25 studies, we found that being male was more susceptible to severe COVID-19. In terms of comorbidities, patients combining diabetes, hypertension, cardiovascular disease and COPD were more likely to develop severe COVID-19, which was consistent with the findings of Guan Wei-jie et al. to some degree. Fever and cough were the main clinical symptoms in both severe and non-severe cases, which was consistent with previous studies[1, 2, 32] . As for complications, ARDS, AKI or shock were much more likely to observed in severe cases, which was in accordance with the finding on MERS-CoV[6, 33]. The death rate of severe cases was obviously higher than the WHO estimates as 2.90%, while it was lower in non-severe cases.
Based on results of clinical symptoms, we found that a significant difference between severe and non-severe patients with COVID-19 on overall factors. But in clinical practice, it is difficult to conclude whether a patient is more likely to develop severe or non-severe COVID-19 based on such clinical symptoms. Nonetheless, clinical symptoms are undoubtedly essential for susceptible cases screening.
According to our analysis results on comorbidities, severe patients used to be with comorbidities on admission especially as diabetes, hypertension and cardiovascular disease, which could affect some key mediators of the host's innate immune response. Previous findings on MERS-CoV also found that people with severe illness were more likely to combine these underlying comorbidities. This can be explained by the phenomenon of cytokine storm that a variety of cytokines gather in the body fluids. Early studies of MERS-CoV found that the amount of Th1/Th2 cytokines profile was higher in patients with diabetes, hypertension or cardiovascular disease which was linked with exacerbation of pro-inflammatory state and generation of oxidative stress[17, 34-38]. Studies have shown that cytokine storm indicate poor prognosis and tissue damage . So far in COVID-19 patients, research has shown that ICU patients had higher plasma levels of IL-2, IL-7, IL-10, GSCF, IP10, MCP1, MIP1A, and TNF-α compared with non-ICU patients.
Considering that these cytokines mainly belong to Th1 or Th2 subgroups, we infer that patients with comorbidities, especially those with diabetes, hypertension or cardiovascular disease, are more likely to develop severe COVID-19. Therefore, we suggest that clinicians can pay more attention to patients with comorbidities, which may prevent the development of severe COVID-19 and its progressive complications with suitable care.
Also, it is believed that cytokine storm is also an important cause of ARDS and multiple organ failure in patients with viral infections[39, 40].
Therefore, we believed that patients combined with diabetes, hypertension or cardiovascular disease were more likely to develop progressive complications, which was fatal and difficult to cure.
As mentioned on complications of severe and non-severe patients, we found that the incidence of ARDS, AKI and shock were remarkably higher in severe patients. This was also consistent with the conclusion of previous research that secondary pneumonia, ARDS, encephalitis, myocarditis and other fatal complications could occur in severe patients[6, 33]. These severe clinical manifestations caused by the underlying comorbidities can also be seen in other respiratory diseases such as influenza and influenza H1N1[32, 39, 41]. With evaluating the occurrence of complications induced by SARS-CoV-2 infection, it helps us fully understand the adverse impact and disease burden of severe COVID-19.
In general, figuring out differences on comorbidities, clinical symptoms and complications between severe and non-severe patients may serves as the superb way to effectively prevent patients who are non-severe but combined comorbidities from developing severe COVID-19. Besides, due to the similarity between COVID-19 with SARS and MERS to a certain extent, we could draw some experience in the previous studies of SARS and MERS while comparing with the studies of COVID-19 as well. We hope that this assessment may aid the public health sector while developing policies for surveillance and response to COVID-19 and its severe outcomes. Also, we hope that it may aid the development of vaccination for those vulnerable people.