Table 4.

Cox proportional hazards model of 235 patients starting dialysis for ESRD using Cockcroft and Gault formula for estimating creatinine clearance (12)a

Variable (Reference Group)Single StepBackwards Logistical Regression
Hazard Ratio95% CISignificanceHazard Ratio95% CISignificance
a Variables entered were either significant on univariate analysis (P < 0.2) or felt to be important, such as gender, weight, and hemoglobin. Variables used were from around the point when estimated creatinine clearance for each patient was 20 ml/min. Survival is timed from eCCr = 20 ml/min to endpoint. Analysis was performed initially as a single step and then as backward logistical regression.
Gender (hazard for male = 1.0)1.550.91 to 2.640.1051.781.09 to 2.910.022
eCCr at start of dialysis (per ml/min)1.101.00 to 1.200.0511.111.01 to 1.210.024
Mode of dialysis (hazard for PD patients = 1.0)1.510.70 to 3.250.29
Initial access via central line (hazard for no = 1.0)1.630.90 to 2.950.1092.181.35 to 3.510.001
Diabetes (hazard for no diabetes = 1.0)1.430.69 to 2.950.330
Albumin (per mg/L)0.990.95 to 1.040.790
White cell count (per 1 × 109/l)1.141.04 to 1.240.0031.131.04 to 1.230.004
Mean arterial pressure (per mmHg)1.011.00 to 1.020.0061.011.00 to 1.020.013
Hemoglobin (per g/dl)0.920.79 to 1.080.320
Wright/Khan index (per point increase)2.461.42 to 4.260.0012.951.92 to 4.54<0.0001
Age (per yr)1.041.02 to 1.060.00041.041.02 to 1.060.0006
Weight (per kg)0.990.98 to 1.010.529