TABLE 4.

Factors independently predicting in-stent restenosis.

Items	Forward stepwise multivariate logistic regression model
	P-value	OR	95% CI
			Lower	Higher
Diabetes mellitus	0.003	5.630	1.766	17.941
Hypercholesteremia	0.013	6.514	1.496	28.363
SUA	0.004	1.010	1.003	1.017
HsCRP	<0.001	1.257	1.123	1.406
Patients with two target lesions	0.005	4.731	1.589	14.084

Factors with P value < 0.05 in univariate logistic regression model were included in this forward stepwise multivariate logistic regression model to screen independent predictors. The predictive model of in-stent restenosis was as follows: P = e^ [−10.322 + 1.728 (diabetes mellitus) + 1.874 (hypercholesteremia) + 0.010 (SUA) + 0.229 (HsCRP) + 1.554 (patients with two target lesions)] / 1 + e^ [−10.322 + 1.728 (diabetes mellitus) + 1.874 (hypercholesteremia) + 0.010 (SUA) + 0.229 (HsCRP) + 1.554 (patients with two target lesions)], −2Ln(L) = 95.594. OR, odds ratio; CI, confidence interval; SUA, serum uric acid; HsCRP, high-sensitivity C-reactive protein.