Table 4.

Cox proportional hazards model for cardiovascular mortality (for comment see also Results)a

ParameterUnits of IncreaseHazard Ratio95% CIP
aIMT-based model: -2 log likelihood = 246.463. Forcing left ventricular mass index (LVMI) into the model produced better data fitting (-2 log likelihood = 238.215; P = 0.005 versus model based on IMT only, see above). In this model, LVMI had an independent prediction power for cardiovascular death. Data are expressed as hazard ratios, 95% confidence interval (CI), and P. Out of the model: treatment modality (P = 0.10), antihypertensive therapy (P = 0.40), homocysteine (P = 0.44), CRP (P = 0.47), systolic BP (P = 0.55), DCCA (P = 0.58), duration of regular dialysis treatment (P = 0.63), hemoglobin (P = 0.70), albumin (P = 0.74), smoking (P = 0.75), fibrinogen (P = 0.79), parathyroid hormone (P = 0.81), LDLcholesterol (P = 0.83), previous cardiovascular events (P = 0.94), total cholesterol (P = 0.94), and diabetes (P = 0.97).
Age1 yr1.051.02 to 1.080.001
Calcium phosphate1 mmol2/L21.651.20 to 2.260.002
Male gender4.531.75 to 11.720.002
Lipoprotein(a)1 mg/dl1.021.01 to 1.030.004
IMT0.1 mm1.241.06 to 1.440.007