TABLE 3.
Summary of the prognostic performance of different severity scores for mortality/intubation in hospitalized patients with COVID-19.
| Risk score | Variables | Risk stratification in original study | OR (95% CI) for mortality, intubation* | RO-AUC for mortality, intubation* |
| Shang COVID severity score (12) | Age, coronary heart disease, Lymphocyte < 8%, Procalcitonin > 0.15 ng/ml, D-dimer > 500 ng/ml | Total score > 2 points defined as high risk; mortality rate 10% vs. 81.1% in low-risk group vs. high-risk group (p < 0.01) | 2.40 (1.72–3.36), 1.76 (1.44–2.15) | 0.836, 0.73 |
| SEIMC (11) | Age, lowest SpO2, NLR, eGFR, dyspnea, sex | 6–8 points defined as high risk, mortality rate was 10.6–19.5%; 9–30 points defined as very high, mortality rate was 27.7–100.0% | 1.17 (1.09–1.25), 1.09 (1.04–1.15) | 0.807, 0.70 |
| Inflammation-based risk score (14) | WBC ≥ 9.3 × 103 cells/μL, CRP level ≥ 13.0 mg/L, serum albumin level ≤ 3.6 g/dl | 5–6 points defined as severe risk, 71% of IMV rate | 1.63 (1.27–2.10), 1.45 (1.24–1.71) | 0.775, 0.69 |
| VICE (17) | DM, SpO2/FiO2, LDH | 4th quintile defined as 0.52–0.81 points, 66.30% of IMV rate 5th quintile defined as 0.81–0.99 points, 90.20% of IMV rate |
51.8 (12.1–221.1), 187.0 (45.1–766.1) | 0.804, 0.82 |
| COVID-IRS-NLR score (13) | Respiratory rate, SaFiO2, LDH, NLR | 5–8 points defined as high risk, 36.6–69.5% of IMV rate 9–11 points defined as very high risk, 90.9–92.8% of IMV rate 12–13 points defined as very high risk, 100% of IMV rate. |
1.51 (1.28–1.78), 1.81 (1.53–2.13) | 0.781, 0.82 |
eGFR, estimated glomerular filtration rate; SpO2, peripheral arterial oxygen saturation; FiO2, fraction of inspired oxygen; SaFiO2, ratio of oxygen saturation to fraction of inspired oxygen; CRP, C-reactive protein; NLR, neutrophil to lymphocyte ratio; DM, diabetes mellitus; CRP, C-reactive protein; LDH, lactic dehydrogenase; IMV, invasive mechanical ventilation; OR, odds ratio; ROC-AUC, receiver-operator characteristic with area under the curve; CI, confidence interval.
*Results of external validation within our cohort.