Table 4.
Univariate and multivariate analysis of death.
| Variables | Univariate analysis | Multivariate analysis | |||
|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | p | ||
| Age (years) | <65 (vs. ≥65) | 2.1 (0.8–5.6) | 0.16 | ||
| Sex | Male (vs. female) | 1.5 (0.6–4.0) | 0.39 | ||
| BMI (kg/m2) | <30 (vs. ≥30) | 3.7 (0.5–27.9) | 0.20 | ||
| Cancer stage | Metastatic (vs. non-metastatic) | 1.2 (0.4–3.4) | 0.74 | ||
| PS | ≥2 (vs. <2) | 4.8 (1.7–13.6) | 0.003 | 3.9 (1.1–13.8) | 0.04 |
| Treatment | Steroids (vs. none) | 2.2 (0.8–5.7) | 0.12 | ||
| Anticoagulants (vs. none) | 0.5 (0.1–3.6) | 0.47 | |||
| Time from symptoms onset (days) | ≥6 (vs. <6) | 0.8 (0.3–2.1) | 0.58 | ||
| Shortness of breath | Yes (vs. no) | 10.2 (2.3–44.8) | 0.002 | ||
| SaO2 | ≤ 93% (vs. >93%) | 6.1 (2.0–18.6) | 0.002 | 1.5 (0.4–6.2) | 0.58 |
| Lymphocytes | <1 ×109/L (vs. ≥1 ×109/L) | 1.1 (0.4–2.9) | 0.92 | ||
| Consolidation | Yes (vs. no) | 3.1 (1.1–9.0) | 0.03 | 3.0 (0.9–10.2) | 0.07 |
| Bilateral distribution | Yes (vs. no) | 1.0 (0.3–3.2) | 0.95 | ||
| Pleural effusion | Yes (vs. no) | 3.7 (1.4–9.6) | 0.007 | 1.5 (0.5–4.7) | 0.54 |
| SARS-CoV-2 RT-PCR | 0.25 | ||||
| Positive (vs. negative) | 2.1 (0.5–9.0) | 0.34 | |||
| Not done (vs. negative) | 0.5 (0.04–5.0) | 0.52 | |||
| C-reactive protein (mg/L) | ≥89 (vs. <89) | 2.2 (0.8–6.0) | 0.15 | ||
| Chest CT extent of COVID-19 pneumonia | <0.001 | 0.01 | |||
| 14–30% (vs. <14%) | 3.9 (1.0–15.2) | 0.047 | 3.4 (0.7–16.9) | 0.14 | |
| ≥30% (vs. <14%) | 18.8 (4.8–73.8) | <0.001 | 12.0 (2.2–64.4) | 0.004 | |
Shortness of breath was excluded from the univariate analysis because of collinearity with SaO2. HR, hazard ratio; CI, confidence interval; BMI, body mass index; PS, performance status; SaO2, oxygen saturation.
Bold p-values were considered statistically significant (p < 0.05).