Distribution of angiographic measures of restenosis after drug-eluting stent implantation

Background: A bimodal distribution of measures of restenosis has been demonstrated at 6-8 months after bare metal stent implantation. Drug-eluting stent (DES) treatment has attenuated the impact of certain factors (eg, diabetes) on restenosis but its effect on the distribution of indices of restenosis is not known. Objective: To perform a detailed analysis of the metrics of restenosis indices after DES implantation. Design, settings, patients: Prospective observational study of patients undergoing DES implantation (Cypher, sirolimus-eluting stent; or Taxus, paclitaxel-eluting stent) at two German centres, with repeat angiography scheduled at 6-8 months after coronary stenting. Main outcome measures: In-stent late luminal loss (LLL) and in-segment percentage diameter stenosis (%DS) as determined by quantitative coronary angiography at recatheterisation. Results: Paired cineangiograms were available for 2057 patients. Overall mean (SD) LLL was 0.31 (0.50) mm; mean (SD) %DS was 30.3 (15.7)%. Distribution of both LLL and %DS differed significantly from normal (Kolmogorov-Smirnov test; p<0.001 for each). For both parameters a mixed distribution model better described the data (likelihood ratio test with 3df; p<0.001 for each). This consisted of two normally distributed subpopulations with means (SD) of 0.10 (0.25) mm and 0.69 (0.60) mm for LLL, and means (SD) of 22.2 (8.6)% and 40.1 (16.6)% for %DS. The results were consistent across subgroups of DES type, "on-label'' versus "off-label'' indication, and presence or absence of diabetes. Conclusions: LLL and %DS at follow-up angiography after DES implantation have a complex mixed distribution that may be accurately represented by a bimodal distribution model. The introduction of DES treatment has not resulted in elimination of a variable propensity to restenosis among subpopulations of patients with stented lesions.