Skip to main content
. 2022 Feb 18;9:811826. doi: 10.3389/fnut.2022.811826

Table 2.

Univariate and multivariate COX regression models for male patients with ACLF.

Univariate Multivariate
Variables HR (95% CI for HR) p.value HR (95% CI for HR) p.value
Age
(years)
1.040
(1.010–1.060)
0.004
Cirrhosis 2.820
(1.490–5.350)
0.002 2.746
(1.229–6.136)
0.014
BMI
(kg/m2)
0.983
(0.918–1.050)
0.615
Obesity 0.734
(0.406–1.330)
0.306
ALB
(g/L)
1.000
(0.956–1.050)
0.915
TB
(mg/dL)
1.050
(1.030–1.070)
0.000 1.041
(1.016–1.067)
0.001
INR 2.390
(1.570–3.630)
0.000 2.800
(1.798–4.363)
0.000
CR
(mg/dL)
1.160
(0.642–2.110)
0.616
Na
(mmol/L)
0.912
(0.866–0.960)
0.000
WBC
(*109/L)
0.987
(0.949–1.030)
0.533
PLT
(*109/L)
0.997
(0.993–1.000)
0.191
HGB
(g/L)
0.992
(0.982–1.000)
0.163
MELD score 1.100
(1.040–1.160)
0.000
Ascites 3.630
(1.450–9.110)
0.006 4.902
(1.152–20.858)
0.031
HE 1.840
(1.060–3.210)
0.031
PMI
(cm2/m2)
0.851
(0.734–0.987)
0.033

The multivariate logistic regression model was fited with a stepwise selection method using statistically baseline factors that had been screened in univariate analysis. BMI, body mass index; ALB, albumin; TB, total bilirubin; INR, International normalized ratio; CR, Serum creatinine; Na, Serum sodium; WBC, White blood cell count; PLT, platelet; HGB, hemoglobin; MELD, model for end-stage liver disease; HE, Hepatic encephalopathy; PMI, psoas muscle index.