Demographic and clinical characteristics of the study population
In total, 239 COVID-19 patients were included in this study. In the present cohort, 216 (90.38%) patients were mild/moderate cases, and 23 (9.62%) cases with progression to severe. Demographic and clinical characteristics of the study population are presented in table 1. The median age was 58 years (range, 26-90 years), and 58.20% (139/239) were male. The median max temperature was 38℃ (range, 36.0-41.0).
When comparing demographic data at admission, progression to severe cases were more likely to be male patients and older age (>75 years) than patients without progression (P<0.05, table 1). The clinical manifestations were mainly presented as the followings (Table 1): fever 77.8% (186/239), cough 60.3% (144/239), expectoration 23.8% (57/239), dyspnea 10.5% (25/239), chest pain 4.6% (11/239), angina 4.2% (11/239), fatigue 28.5% (68/239), myalgia 9.2% (22/239), headache 5.0% (12/239), vomit 1.7% (4/239) and diarrhea 18% (43/239).
The clinical characteristics of the study population were summarized in table 2. When comparing biochemical indexes of COVID-19 between moderate cases with and without progression, we found that there were significant differences in lymphocytes, NLR, CRP, AST, TBIL, DBIL, Cr, urea, glucose, sodium, PT, CD3+ T cell, CD4+ T cell, CD8+ T cell, CD19+ T cell, CD16+/ CD56+ NK cell between patients with or without progression (all P<0.05), detailed information were listed in Table 2.
Univariant and multivariant COX regression model for progression from mild/moderate cases into severe cases
When exploring risk factors of progression from mild/moderate cases into severe COVID-19 cases, we compared the demographic and clinical data between moderate cases and progression to sever cases. Using univariant and multivariant COX regression model, the results showed a significant difference in pulmonary disease (11.20, 95%CI 2.50-49.70, p=0.001), age over 75 (3.92, 95%CI 1.61-9.73, p=0.003), IgM (6.31, 95%CI 1.99-19.60, p=0.002) , CD16+/CD56+ NK cell (3.40, 95%CI 1.31-9.13, p=0.014) and AST(4.60, 95%CI 1.31-16.00, p=0.018) (Table 3), which were the 5 independent risk factors for progression from mild/moderate cases into severe cases (Fig. 1). However, there were no significant impacts of other variables showed in our study population (see Supplementary material, Table S1). We also used global Schoenfeld test cox diagnostics deviance and ox proportional hazards model fit to evaluate these five independent risk factors for progression from mild/moderate cases into severe cases, suggesting good performances (see Supplementary material, Fig. S1-S3).
Moreover, the Kaplan-Meier survival curve analysis and log-rank test showed a significant difference in survival curve in COVID-19 patients categorized by the Pulmonary disease, Age, IgM, CD16+/CD56+ NK cells and AST, respectively (see Supplementary material, Fig. S4 a-e).
Development of predictive score for progression from moderate cases into severe cases
Predictors including pulmonary disease, Age, IgM, CD16+/CD56+ NK cells and AST were enrolled in the development of predictive score for COVID-19 patients’ progression from mild/moderate cases into severe cases. The new predictive score (Pulmonary disease, Age, IgM, CD16+/CD56+ NK cell, AST; PAINT score) = (pulmonary disease) *2.4174 + (Age>75) *1.3594 + (IgM<0.84) *1.8399 + (CD16+/CD56+NK cell<116.5) *1.2246 + (AST>25) *1.5182.
The points contributed for each variable are shown in figure S5. To demonstrate the ability of the newly predictive score to find more severe patients for early clinical treatment, Kaplan-Meier survival curve analysis was used to find the best cut-off value; a value of 14.687 points was found to divide the patients into mild/moderate and progression to sever groups (P = 0.001, Fig. 2).
We performed ROC analysis to evaluate the efficacy of PAINT score model for predicting COVID-19 patients’ progression from mild/moderate cases into severe cases. We compared the PAINT score with qSOFA and CURB-65 ((confusion, uraemia, respiratory rate, BP, age>65 years)) score. As demonstrated in fig. 3, the C-index of the newly predictive progression model for predicting progression from mild/moderate cases into severe cases was 0.902 ± 0.021. However, the C-index of qSOFA and CURB-65 score for the prediction of progression was 0.534 ± 0.027 and 0.561 ± 0.058. We also compare the newly predictive progression model with the 5 the independent risk factors (pulmonary disease, Age, IgM, CD16+/CD56+ NK cells and AST), the C-index for the prediction of progression was 0.5432 ± 0.034, 0.639 ± 0.052, 0.683 ± 0.044, 0.647 ± 0.050 and 0.716 ± 0.036,respectively (see Supplementary material, Table S2).These findings suggested that the PAINT score might be suitable for predicting progression from mild/moderate cases into severe cases.
For internal validation of the ability of the newly predictive progression model, we performed concordance index analysis to evaluate the discrimination for the PAINT score. The better discrimination was observed in our PAINT score than qSOFA and CURB-65 score (Fig. 4a). Moreover, we performed 1000 times bootstrap internal validations, our newly predictive PAINT score also showed better discrimination (Fig. 4b)
Nomogram, calibration, decision curve and clinical impact curve for progression from mild/moderate cases into severe cases
In our study population, we used 5 variables (Pulmonary disease, Age, IgM, CD16+/CD56+ NK cells and AST) to predict 28 days progression from mild/moderate cases into severe cases. According to the construction principle of nomogram score, each variable will be given different points and weights. The nomogram score was shown in Fig. 5a. We can evaluate the score of each variable in turn according to their clinical characteristics and examination results, and then summarize according to the total score of 5 variables. Based on the total score, the patient can determine the probability of progression to severe COVID-19 cases. The calibration curves for 28 days progression was also performed well in the internal validation set (Fig. 5b). Nomograms and decision curves analyses were also performed well in the PAINT score (Fig. 5c). Clinical impact curves were proposed to assess the clinical usefulness of the risk prediction nomogram (Fig. 5d).