5 Recommendations for further research

5.1 Ongoing trials for paclitaxel-eluting stents include TAXUS I (follow-up of a small initial study of slow-release formulation versus BMS in patients with either previously-untreated lesions or restenosis), TAXUS II (follow-up of a larger study of both slow- and moderate-release formulations versus BMS in patients with previously-untreated lesions), TAXUS IV (a large trial of slow-release formulation versus BMS stratified by presence or absence of diabetes and by vessel diameter), TAXUS V (focussing on small vessels, long lesions, bifurcations and in-stent restenosis) and TAXUS VI (moderate release for long lesions). For sirolimus-eluting stents, on-going trials include RAVEL (small-diameter vessels), SIRIUS (high risk for cardiovascular disease progression and restenosis due to the diabetes, exposure to multiple stent implantation and use of overlapping stents) and E-SIRIUS (previously-untreated single vessels of diameter 2.5 to 3 mm and for lesions between 15 and 32 mm in length); and FUTURE (previously-untreated vessels between 2.75 and 4 mm, less than 28 mm long ) for everolimus-eluting stents. REALITY, a head-to-head trial of the Cypher sirolimus DES and Taxus paclitaxel DES, is under way.

5.2 Until now, trials have been restricted to single-artery studies for the sake of simplicity and ease of interpretation. Extrapolation of results to more than one artery critically depends on untested assumptions. Randomised controlled trials (RCTs) of the use of DES in more than one artery concurrently are therefore required, in order to confirm or refute the appropriateness of the extrapolations used in the modelling.

5.3 To compare long-term outcomes, particularly with respect to stents against CABG, much longer trial follow-ups are required.

5.4 New BMS designs should be tested against current BMS and DES designs.

5.5 Head-to-head RCTs of those DES that have been CE marked and have been shown to be clinically superior to the corresponding BMS are required.

5.6 Studies to determine whether diabetes is a risk factor for increased rate of restenosis following PCI, independent of lesion length and artery calibre, are required. Much of this work could be performed by an analysis of patient-level data taken from trials already conducted.