78 STENTED LENGTH-TO-DIAMETER RATIO IS STRONGLY ASSOCIATED WITH HIGHER COMPLEXITY OF IN-STENT RESTENOSIS, INDEPENDENT OF DIABETES AND CONVENTIONAL ANGIOGRAPHIC PREDICTORS

• Published in: Journal of investigative medicine 2005-03, Vol.53 (2), p.S400-S400
• Main Authors: Nathan, S., Chugh, A. R., Amin, A. P., Al-Dallow, R., Haynie, J., Almeda, F. Q.
• Format: Article
• Language: English
• ISSN: 1081-5589; 1708-8267
• DOI: 10.2310/6650.2005.00205.77

BackgroundIn-stent restenosis (ISR) of bare-metal stents (BMS) occurs in a significant proportion of stented patients and has been associated with specific clinical characteristics such as diabetes mellitus (DM), lesion length (LL) and smaller minimal luminal diameter (MLD). An angiographic grading system was previously developed by Mehran et al and denotes increasing ISR lesion complexity. Increased clinical events have been shown to parallel increased ISR lesion complexity. It is unclear if increasing ISR complexity is associated with DM when adjusted for LL and MLD and if the ratio of stent length to diameter (L/D) is more predictive of ISR lesion complexity than either measurement alone.MethodsMultiple data from 199 patients (78 DM, 121 non-DM) presenting with BMS ISR to 2 institutions were collected. ISR lesion morphology (Types I-IV) was adjudicated by angiographers. Stent dimensions were obtained by averaged measurements using GeQCA 2.0 software. Relationships of clinical variables to ISR complexity were evaluated via univariate analysis using ANOVA and Pearson's chi-square. Univariate predictors and a priori hypothesis variables were entered into a multivariate logistic regression model. Lesion complexity was dichotomized into Type I (simple) vs. Types II-IV (complex) to conserve power. Receiver operator characteristic (ROC) curves were constructed for stent L/D ratio and LL and MLD individually, for DM and non-DM patients. Areas (AUC) for the ROC curves were compared using standard errors to determine the single most predictive parameter.ResultsBy multivariate regression, DM [OR 2.22, 95% CI 1.13 - 4.38, p = 0.021] and L/D ratio [OR 1.27, 95% CI 1.11-1.46, p = 0.001] were independently associated with diffuse ISR. ROC curve analysis suggested L/D ratio [AUC 0.71, 95% CI 0.63-0.78] confers an increase in predictive power over LL alone [AUC 0.69, 95% CI 0.62-0.77, p = 0.60] and is associated with a higher risk for complex ISR than LL [OR 1.10, 95% CI 1.05 - 1.16, p