Patients
Between February 8 to 18, 2020, 138 patients with diagnosed COVID-19 were consecutively admitted. All patients met the inclusive/exclusion criteria and were included in the analysis. Among the enrolled patients, 119 were diagnoses as severe cases and 16 as critically ill cases at hospital admission; during hospital stay, 19 more severe cases progressed to critical illness, resulting a total of 35 critically ill cases (Fig 1).
Of all enrolled patients, the age was mean 62 (SD 14) years, and 51.4% (71/138) were male. Symptoms that appeared in more than half of patients included fever (86.2%), cough (84.1%), dyspnea (66·7%), and expectoration (62.0%). The median time from symptom onset to hospital admission was 14 (interquartile range [IQR] 11–18) days. The median time from diagnosis to hospital admission was 8 (IQR 4–13) days. Before admission, 71.0% (98/138) of patients received antiviral drugs and 68.1% (94/138) received antibiotics (Table 1).
Table 1. Baseline characteristics (see at the end of this text)
Baseline characteristics and variables at admission
Of all enrolled patients, when compared with severe cases without aggravation, those who developed critical illness were older, and received fewer antiviral drugs (especially arbidol) and antibiotics (especially fluoroquinolones) before admission (Table 1). At hospital admission, patients who developed critical illness had a faster respiratory rate, a greater proportion with pulse oxygen saturation <93%, higher white blood cell and neutrophil counts, but a lower lymphocyte count; regarding biochemical test results, they had higher serum levels of aspartate aminotransferase, creatinine, blood urea nitrogen, lactate dehydrogenase, myoglobin, hypersensitive cardiac troponin I, creatine kinase-MB, and N-terminal pro-brain natriuretic peptide, but a lower level of serum albumin; regarding coagulation, they had a longer prothrombin time and a higher D-dimer level (Table 2; Additional file 1: Table S1).
Table 2. Variables at admission (see at the end of this text)
In the subgroup of severe cases at hospital admission, when compared with cases without aggravation, those who developed critical illness received fewer arbidol before admission (Table 1). At hospital admission, patients who developed critical illness had a faster respiratory rate, a greater proportion with pulse oxygen saturation <93%, higher white blood cell and neutrophil counts, but a lower lymphocyte count; regarding biochemical test results, they had higher serum levels of aspartate aminotransferase, blood urea nitrogen, lactate dehydrogenase, myoglobin, hypersensitive cardiac troponin I, creatine kinase-MB, and N-terminal pro-brain natriuretic peptide; regarding coagulation, they had a longer prothrombin time, a longer activated partial thromboplastin time, and a higher D-dimer level (Table 2; Additional file 1: Table S1).
Treatments after admission and outcomes
Of all enrolled patients, when compared with severe cases without aggravation, those who developed critical illness received more antibiotics (especially fluoroquinolones), more glucocorticoids, more gamma globulin, and more noninvasive/invasive ventilation; they developed critical illness in a median 2 (IQR 0-8) days, and had a higher in-hospital mortality. Those who survived had a longer hospital stay (Table 3).
Table 3. Treatments after admission and outcomes (see at the end of this text)
In the subgroup of severe cases at hospital admission, when compared with cases without aggravation, those who developed critical illness received more antibiotics, more glucocorticoids, and more noninvasive/invasive ventilation; they developed critical illness in a median 5 (IQR 4-11) days and had a higher in-hospital mortality (Table 3).
Predictors of progressionfrom severe type to critically ill type
Of all enrolled patients, univariable analysis identified 26 factors with P ≤0.10 (Additional file 2: Table S2). After excluding factors having collinearity or clinical relation with others, 13 factors were included in the multivariate logistic regression model. Five factors were identified to be independently associated with the development of critical illness in COVID-19 patients; of them longer duration from diagnosis to admission (odds ratio [OR] 1.108, 95% CI 1.022-1.202; P=0.013), pulse oxygen saturation at admission <93% (OR 5.775, 95% CI 1.257-26.535; P=0.024), higher neutrophil count (OR 1.495, 95% CI 1.177-1.899; P=0.001) and higher creatine kinase-MB level (OR 2.449, 95% CI 1.089-5.511; P=0.030) at admission were associated with a higher risk, whereas higher lymphocyte count at admission (OR 0.149, 95% CI 0.026-0.852; P=0.032) was associated with a lower risk of critical illness development (Table 4).
Table 4. Predictors of progression to critically ill type (see at the end of this text)
In the subgroup of severe cases at hospital admission, univariable analysis identified 16 factors with P ≤0.10 (Additional file 2: Table S2). After excluding factors having collinearity or clinical relation with others, 11 factors were included in the multivariate logistic regression model. Four factors were identified to be independently associated with the development of critical illness in COVID-19 patients; of them longer duration from diagnosis to admission (OR 1.085, 95% CI 1.009-1.167; P=0.027), pulse oxygen saturation at admission <93% (OR 11.182, 95% CI 2.426-51.534; P=0.002) and higher neutrophil count at admission (OR 1.403, 95% CI 1.117-1.763; P=0.004) were associated with a higher risk, whereas higher lymphocyte count at admission (OR 0.147, 95% CI 0.028-0.760; P=0.022) was associated with a lower risk of critical illness development (Table 4).