Development of a new Cox model for predicting long-term survival in hepatitis cirrhosis patients underwent transjugular intrahepatic portosystemic shunts

Table 2 Univariate and multivariate Cox proportional hazards regression models for prognostic factors after transjugular intrahepatic portosystemic shunt placement

	Univariable model				Multivariable model
	HR		95%CI	P value	HR		95%CI	P value
Etiology	0.29	1.34	0.60-2.98	0.48
Indication
Variceal bleeding
Ascites	1.36	3.90	2.01-7.58	< 0.001	1.45	4.27	2.04-8.94	< 0.001b
Both	1.01	2.76	1.34-5.65	0.006	1.00	2.72	1.24-5.97	0.01b
Shunt	1.20	3.33	1.87-5.94	< 0.001	1.83	6.26	2.67-14.63	< 0.001b
Age	-0.02	0.99	0.96-1.01	0.18
Pre-PPG	0.10	1.10	1.05-1.15	< 0.001
Post-PPG	0.25	1.29	1.21-1.37	< 0.001	0.27	1.32	1.22-1.42	< 0.001b
PVP decrease	-0.09	0.92	0.89-0.94	< 0.001	-0.04	0.96	0.93-1.00	0.03a
ALT	-0.001	1.00	0.99-1.01	0.81
AST	0.0008	1.00	0.99-1.01	0.83
ALB	-0.004	1.00	0.94-1.06	0.88
Na	-0.05	0.95	0.89-1.02	0.13
ICGR15	0.04	1.04	1.03-1.06	< 0.001	0.05	1.05	1.03-1.07	< 0.001b
1/INR	-2.99	0.05	0.003-0.77	0.03
Log10 (Cr)	0.46	1.58	0.20-12.50	0.67
Log10 (Tbil)	2.95	19.16	4.87-75.41	< 0.001	2.24	9.38	2.16-40.82	0.002b
Log10 (WBC)	-0.43	0.65	0.21-1.99	0.45
Log10 (PLT)	-1.17	0.31	0.12-0.78	0.01
Log10 (HGB)	1.94	6.98	0.52-92.97	0.14

^aP < 0.05.

^bP < 0.01.

As the Cr, Tbil, WBC, HGB and PLT data showed highly skewed distributions, logarithmic transformations were performed. The INR was also highly skewed, and inverse transformation was used to achieve a sufficient normal distribution. ALB: Albumin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; Cr: Creatinine; HGB: Hemoglobin; ICGR15: Indocyanine green retention rate at 15 min; INR: International normalized ratio; Na: Sodium; PLT: Platelet; PPG: Portal venous pressure gradient; PVP: Portal venous pressure; Tbil: Total bilirubin; WBC: White blood cell.