(A) Data regarding clinically-driven target lesion revascularization (TLR), stent thrombosis (ST), based on definite or probable cases (as a reliable approximation of the true incidence of stent thrombosis) according to the Academic Research Consortium (ARC) definition and includes events post-TLR (secondary thrombosis), death (D) or myocardial infarction (MI) are shown for sirolimus-eluting stents (SES) and bare-metal stents (BMS). There were 3 additional cases of stent thrombosis in the SES and 10 in the BMS group resulting in a total of 13 versus 15 cases (compared to 10 vs. 5 cases using the protocol-defined criterion). Six out of the 10 additional thrombotic episodes in the BMS group occurred after intervening TLR (5 related to brachytherapy) compared to none in the SES group. There were a total of 100 death or MI events in the SES group (11.4%) compared to 89 in the BMS group (10.2%). (B) Attributable cause of death or myocardial infarction following stenting. Differences in the rates of all-cause death or myocardial infarction (top panel) and all-cause death (bottom panel) between SES (n = 878) and BMS (n = 870) attributable to clinically-driven target lesion revascularization (post-TLR), stent thrombosis (post-ST) based on definite or probable cases according to ARC definition, and other (non-ST or TLR). Data are shown as relative risk and 95% confidence intervals.

The double arrows indicate reversible state-to-state transitions and the single arrows indicate irreversible transitions. The rate of each transition can be quantified in terms of an empirical rate constant or transition probability (not illustrated). Such models can be used to predict the time-dependent prevalence of each state over relatively long periods of follow-up.