Restenosis rates following Palmaz-Schatz stenting vary between 13 and 36%. To determine whether the nearly three-fold interseries difference in restenosis rates reflects systematic factors (demographic and procedural variables such as post-treatment lumen diameter) rather than random variation, we pooled 486 lesions treated with the Palmaz-Schatz coronary stent in the U.S. Multicenter Registry (n = 259) with the single-center Beth Israel Hospital experience (n = 227) and analyzed the combined series using a continuous geometric model of restenosis. A greater proportion of lesions (66 vs. 50%; P < 0.001) in the Multicenter cohort represented restenosis after a prior intervention. The Multicenter experience also showed a smaller minimum lumen diameter following stent placement (2.61 vs. 3.43 mm; P < 0.001), higher post-stent percent stenosis (16 vs. -2%; P < 0.001), and less acute gain (1.79 vs. 2.67 mm; P < 0.001) than lesions treated in the Beth Israel cohort. At 6-month follow-up, the angiographic restenosis rate (> or = 50% diameter stenosis) was significantly higher in the Multicenter group (35 vs. 27%; P = 0.05), despite the development of less absolute late loss (0.96 vs. 1.32 mm; P < 0.001). Multivariable modeling, however, showed that the only independent predictors of restenosis were a post-procedure lumen diameter < 2.80 mm (odds ratio = 1.57; P = 0.04), diabetes mellitus (odds ratio = 3.55; P < 0.001), and prior restenosis (odds ratio = 1.84; P = 0.008).(ABSTRACT TRUNCATED AT 250 WORDS)