Lesion-to-lesion independence of restenosis after treatment by conventional angioplasty, stenting, or directional atherectomy. Validation of lesion-based restenosis analysis.

BACKGROUND Since many restenosis trials include patients in whom more than one lesion is treated, analysis of the angiographic data on a "per lesion" basis might be confounded by potential correlations of restenosis among multiple treated lesions within each patient. The goals of this study were: 1) to determine whether there was any correlation in the rate of restenosis among multiple lesions that underwent conventional angioplasty, stenting, or directional atherectomy within the same patient and 2) to determine whether lesions treated in a multilesion intervention experience a different magnitude of restenosis than lesions undergoing single-lesion procedures. METHODS AND RESULTS Of 441 patients treated by Palmaz-Schatz stenting (n = 114), directional atherectomy (n = 100), or conventional balloon angioplasty (n = 227), 67 underwent multilesion procedures involving treatment of 146 lesions. A general linear model with intraclass correlation (GLIMIC) was used to calculate the coefficient of correlation (rho) of the change in the measured minimum luminal diameter (late loss) from the time of the initial procedure to 6-month angiogram among the multiple lesions within the same patient for all 441 patients. This showed no correlation among multiple lesions within the same patient for the late loss in minimum luminal diameter (rho = -0.12 [95% CI: -0.40, 0.12]), among lesions in the same vessel (rho = 0.14 [95% CI: -0.34, 0.62]), or among different vessels (rho = -0.18 [95% CI: -0.52, 0.16]), suggesting that the magnitude of late loss is independent among multiple lesions within the same patient. There was no difference (p = 0.96) between the observed incidence of zero-, one-, and two-vessel restenosis (> or = 50% diameter stenosis at follow-up) for patients with multiple-lesion treatment and that predicted assuming lesion-to-lesion independence. Similarly, there was no difference in late loss or in the overall binary restenosis rate when single-lesion procedures were compared with multilesion procedures. Multivariable analysis of the late loss in lumen diameter (which adjusted for the effects of the acute result and the device used) demonstrated no independent effect (p = 0.20) of single-lesion versus multilesion status. CONCLUSIONS Luminal encroachment appears to occur at independent rates among multiple lesions treated in a single patient. The observed incidence of restenosis for patients with multiple treated lesions is accurately predicted assuming independent probabilities of restenosis. Lesion-based analysis, even when including multiple treated lesions within the same patient, is thus valid for evaluating conventional angioplasty, stenting, or directional atherectomy.