Table 2.

Univariate and multivariate Cox regression models of all-cause and cardiovascular mortality for AGR groups (all patients = 554).

	All-cause mortality		CVD mortality
	Hazard ratio (95% CI)	p-value	Hazard ratio (95% CI)	p-value
(A) AGR median ( AGR < 1.2 vs. AGR ≥ 1.2)
Univariate model	4.16 (2.79, 6.2)	< 0.001	4.52 (2.69, 7.58)	< 0.001
Model 1	2.46 (1.63, 3.73)	< 0.001	2.76 (1.62, 4.72)	< 0.001
Model 2	2.64 (1.74, 4.02)	< 0.001	2.96 (1.73, 5.09)	< 0.001
Model 3	2.44 (1.59, 3.73)	< 0.001	2.69 (1.56, 4.64)	< 0.001
Model 4	2.32 (1.51, 3.58)	< 0.001	2.51 (1.44, 4.39)	0.001
Model 5	2.12 (1.34, 3.35)	0.001	2.58 (1.42, 4.70)	0.002
(B) Sensitivity tests
(i) AGR as a continuous variable (per 1-unit decrease)	2.89 (1.79, 4.64)	< 0.001	3.27 (1.82, 5.88)	< 0.001
(ii) Optimal AGR by ROC analysis	2.13 (1.43, 3.17)	< 0.001	2.12 (1.28, 3.51)	0.004
(iii) AGR tertiles
Third tertile	1		1
Second tertile	1.68 (0.99, 2.84)	0.052	2.32 (1.2, 4.5)	0.012
First tertile	2.57 (1.57, 4.21)	< 0.001	2.84 (1.49, 5.41)	0.001
P for trend	–	< 0.001	–	0.002
(iv) AGR median adjusted for propensity score	1.74 (1.13, 2.66)	0.012	1.81 (1.04, 3.14)	0.037

Model 1: AGR median, age, sex, BMI and smoking status.

Model 2: Model 1 plus medications.

Model 3: Model 2 plus comorbidities.

Model 4: Model 3 plus peritoneal dialysis-related parameters.

Model 5: model 4 plus laboratory data.

AGR, albumin–globulin ratio.