[Drug-eluting stents for diabetic patients. A critical appraisal of the currently available data from randomized trials]

Abstract

Background and purpose: Patients with coronary artery disease (CAD) and diabetes mellitus represent a peculiar high-risk population because of their specific characteristics of atherosclerotic disease. In conjunction with the diabetes-related comorbidities, percutaneous coronary intervention (PCI) often leads not only to a worse acute result but - also as compared to nondiabetics - to significantly worse long-term results due to the higher restenosis rates. The rapid introduction of effective drug-eluting stents (DES), which undoubtedly reduce the restenosis rates as compared with bare-metal stents (BMS), brought great hope of providing diabetic patients better and longer-lasting interventional solutions. This overview compiles the currently available data from randomized trials and meta-analyses.

Methods: Altogether, 86 randomized DES studies were identified in 34,677 patients. None of these with an adequate primary clinical endpoint had diabetes as an inclusion criterion. The high standard of an adequate primary clinical endpoint for comparing BMS versus DES in patients with predominantly stable CAD and de novo lesions was met by only five pivotal trials: SIRIUS (Cypher stent), TAXUS-IV, -V, -VI (Taxus stent) and ENDEAVOR II (Endeavor stent). Patients with diabetes were represented in these trials only in smaller subgroups. Three studies with a primary surrogate endpoint had diabetes as an inclusion criterion (DIABETES, SCORPIUS and ISAR-DIABETES, totaling 610 patients). The efficacy parameter TLR (target lesion revascularization) was chosen to compare these studies and the subgroup analyses.

Results: In the subgroup analyses for diabetic patients in the pivotal trials, TLR was reduced at the time of the primary endpoint after 9 months as compared with BMS as follows (p < 0.05): Cypher stent (279 patients): 22.3% versus 6.9%; Taxus stent (318 patients): 16.0% versus 5.2%; Endeavor stent (239 patients): 15.2% versus 7.5%. 5-year follow-up data are available only for the Cypher and Taxus stents with 33.1% versus 13.7% and 26.9% versus 13.4%. As compared with BMS, the three above-mentioned studies studies with a primary surrogate endpoint and diabetes as an inclusion criterion showed a significant reduction of TLR after 9 months of 31.3% versus 7.3% and after 8 months of 25.0% versus 5.3%, respectively. The TLR between Taxus (12.0%) and Cypher (6.4%), however, was not statistically different. Meta-analyses of 627 Cypher patients and 814 Taxus patients corroborated these findings with 1-year TLR of 24.8% versus 7.9% and 20.5% versus 8.0%.

Conclusion: Of the 22 DES having received a CE certificate, long-term data over 5 years for patients with diabetes are available only for the Cypher and the Taxus stents. Compared with BMS, patients with diabetes and their characteristically small vessels and long lesions predominantly benefit from effective DES. The sometimes postulated differences between Cypher and Taxus in diabetic patients could not be convincingly demonstrated; larger randomized trials with a primary clinical endpoint are required for this. PCI cannot be considered a scientifically sound and evidence- based alternative to bypass surgery in diabetic patients with multivessel disease and/or unprotected left main stenosis until we have the results of the SYNTAX, COMBAT and FREEDOM trials.