Author + information
Studies have shown that first generation drug-eluting stents (DES) substantially decrease risk of restenosis compared to contemporary bare-metal stents. However, long-term comparisons between DES and BMS are not well-studied. The aim of this study was to compare two stents that were previously frequently used in regard to long-term risk of restenosis and stent thrombosis.
We used data from the SCAAR registry (Swedish Coronary Angiography and Angioplasty Registry) for the PCI procedures performed in V?stra G?taland County in Sweden. The database contains information about all procedures performed at five PCI centers (∼20% of all SCAAR data). All consecutive procedures performed between 2004 and 2014 for stable angina, UA/NSTEMI and STEMI were included in the analysis. We compared two different stents: Cordis Cypher Select (C-CS), and Boston Scientific Liberte (BS-L). The two stents used during the study period are the result of biannual stent procurement procedures in our region. We modelled our data with multilevel Cox proportional-hazards regression with landmark analysis at one-year. The assumption of proportionality of hazards was fulfilled before and after one year of follow-up. We used stents as the primary observational unit with patients and hospitals as random effect variables. To adjust for differences in patient's characteristics the following variables were used: age, gender, hypertension, hyperlipidemia, smoking status, diabetes, severity of coronary artery disease, indication for PCI (stable angina, UA/NSTEMI and STEMI), stent length and stent diameter. The primary combined endpoint was time to first occurrence of either stent thrombosis or restenosis.
During the study period 2210 C-CS and 6941 B-SL were implanted in 5314 patients. The mean follow-up time was 2288 days for C-CS and 2297 days for BS-L. There were 1038 events of which 157 (15.1%) were stent thromboses. Treatment with C-CS was associated with lower risk for restenosis or stent thrombosis up to one year from index procedure (HR 0.41; 95% CI 0.32-0.52; P<0.001). However, after one-year of follow-up, the risk was substantially higher in C-CS (HR 2.81; 95% CI 2.25 - 3.50; P<0.001).
In this observational study, treatment with C-CS-a first-generation DES-was not associated with better outcome than BS-L-a contemporary BMS. Continuation of restenosis and stent thrombosis events long after the index procedure with C-CS stent are of major concern for patient safety.