Table 4.
Variable | Univariate analysis |
Multivariate analysis |
||||
---|---|---|---|---|---|---|
OR | 95% (CI) | p value | OR | 95% (CI) | p value | |
Gastrointestinal bleeding | ||||||
Age | 0.98 | 0.95–1.0 | 0.15 | |||
Sex, male | 1.1 | 0.55–2.5 | 0.75 | |||
Liver cirrhosis | 2.1 | 0.99–5.0 | 0.07 | |||
CPS B (CPS A as reference) | 1.3 | 0.52–3.0 | 0.54 | |||
CPS C (CPS A as reference) | 1.6 | 0.34–5.5 | 0.51 | |||
MVI | 0.96 | 0.44–2.0 | 0.91 | |||
Platelets | 1 | 0.99–1.0 | 0.14 | |||
Spleen size | 1.2 | 1.0–1.3 | 0.007 | 1.1 | 1.0–1.3 | 0.03 |
Presence of varices | 1.8 | 0.91–3.5 | 0.09 | |||
History of variceal bleeding | 3.9 | 1.4–10 | 0.007 | 3.0 | 1.0–8.4 | 0.04 |
NSBB | 0.8 | 0.37–1.6 | 0.54 | |||
Prior EBL | 2.1 | 0.91–4.5 | 0.07 | |||
Anticoagulation | 1.5 | 0.75–2.8 | 0.26 | |||
Antiplatelets |
0.76 |
0.34–1.6 |
0.47 |
|||
Variceal bleeding | ||||||
Age | 0.96 | 0.92–1.0 | 0.07 | |||
Sex, male | 2.2 | 0.59–14 | 0.31 | |||
CPS B (CPS A as reference) | 2.4 | 0.70–7.1 | 0.14 | |||
CPS C (CPS A as reference) | 1.4 | 0.072–8.1 | 0.77 | |||
MVI | 1.6 | 0.51–5.1 | 0.40 | |||
Platelets | 1 | 0.99–1.0 | 0.18 | |||
Spleen size | 1.2 | 1.0–1.4 | 0.04 | |||
Presence of varices | 4.5 | 1.4–21 | 0.02 | |||
Esophageal varices grade I | 2.7 | 0.59–14 | 0.2 | |||
Esophageal varices grades II–III | 8.6 | 2.2–42 | 0.003 | |||
History of variceal bleeding | 2.5 | 0.37–9.8 | 0.26 | |||
NSBB | 1.2 | 0.38–3.5 | 0.7 | |||
Prior EBL | 2.4 | 0.64–7.2 | 0.15 | |||
Anticoagulation |
2 |
0.68–5.4 |
0.2 |
|||
Non-gastrointestinal bleeding | ||||||
Age | 0.99 | 0.95–1.0 | 0.53 | 1 | 0.97–1.0 | 0.74 |
Sex, male | 2.2 | 0.84–7.7 | 0.14 | 2.1 | 0.79–7.4 | 0.17 |
Liver cirrhosis | 2.3 | 0.92–6.9 | 0.10 | |||
CPS B/C (CPS A as reference) | 0.24 | 0.037–0.83 | 0.05 ns | |||
Platelets | 1 | 1.0–1.0 | 0.57 | |||
Anticoagulation | 2.3 | 1.1–4.9 | 0.02 | 2.2 | 1.0–4.7 | 0.04 |
Antiplatelets | 1.5 | 0.65–4.3 | 0.36 |
ORs for the risk of bleeding were calculated using both univariate and multivariate logistic regression models. Liver cirrhosis was not included in the model for risk factors of variceal bleeding, because the presence of liver cirrhosis and the occurrence of variceal bleeding were linearly dependent predictors causing estimation instability unsuitable for logistic regression analysis. The bold emphasis indicate statistical significant results. azeto/bev, atezolizumab/bevacizumab; CPS, Child–Pugh score; EBL, endoscopic band ligation; MVI, macrovascular invasion; NSBB, non-selective beta-blocker; OR, odds ratio.