Sequence Variants of Peroxisome Proliferator-Activated Receptor-Gamma Gene and the Clinical Courses of Patients with End-Stage Renal Disease

Table 4

The predictors of ESRD identified by fitting multiple logistic regression model weighted by the inverse of survival probability using the stepwise variable selection method.

Covariate^*	Regression coefficient	Standard error	t value	P value	Odds ratio	95% confidence interval of odds ratio
Intercept	0.075	0.192	0.390	0.697
Age (years)	−0.008	0.003	−2.666	0.008	0.992	0.985	—	0.998
DM	1.811	0.141	12.846	<0.001	6.225	4.724	—	8.203
Non-DM × PPAR-γ Pro12Ala-GG	13.939	0.511	27.266	<0.001			—

Multiple weighted logistic regression model: ; the estimated area under the receiver operating characteristic (ROC) curve = 0.689.
DM: diabetes mellitus; ESRD: end-stage renal disease.
^*The symbol “×” indicates the interaction between two covariates and it can be literally interpreted as “and” in this case.