Table 5.

OPG (ln) as the dependent variable and correlations with potential independent variables (univariate and multiple linear regression analyses)a

Independent VariablesLinear RegressionMultiple Linear Regression
βCorrelation (r)PβSE (β )P
Creatinine clearance−0.0100.45<0.001−0.0040.0020.006
hsCRP (ln)0.1010.35<0.0010.0690.019<0.001
CsA (daily dose)−0.0010.32<0.001−0.00030.00020.165
New-onset PTDM0.2450.240.0010.0330.0660.613
Active CMV infection0.1870.230.0030.0110.0590.851
Pretransplantation CVD0.1970.180.018−0.0690.0770.366
Total steroid dose0.0530.160.0330.0270.0220.210
Nephrosclerosis (cause renal failure)0.1880.150.0450.1140.0800.158
  • a The multiple linear regression model explains 42% of the variation in serum OPG (ln); R2 = 0.42. The partial correlation coefficients/squared partial correlations for age = 0.321/0.10, creatinine clearance = −0.219/0.05, and hsCRP = 0.278/0.08 suggest that these variables independently explain 10, 5, and 8% of the variation in OPG (ln), respectively. hsCRP, high-sensitivity C-reactive protein; PTDM, posttransplantation diabetes; CVD, cardiovascular disease.