In October 2005, in a late breaking trial session at that year’s TCT meeting, Dr. David Holmes (Mayo Clinic, Rochester, MN), lead investigator of the Sirolimus-Eluting Stent versus Intravascular Brachytherapy in the Treatment of Patients with In-Stent Restenotic Coronary Lesions (SISR) trial, reported superiority of the Cypher sirolimus-eluting stent (SES) (Cordis, Miami Lakes, FL) over vascular brachytherapy (VBT) for the treatment of bare metal stent (BMS) in-stent restenosis (ISR).

The SISR study was a multicenter, randomized trial of 384 patients who presented with ISR of native coronary lesions. Patients were randomized in a 2:1 manner to SES or VBT (beta or gamma radiation). In an interview given to theheart.org, Dr. Holmes stated the following, “We can say that both treatments, vascular brachytherapy and the sirolimus stent, are effective at reducing neointimal hyperplasia within the treated lesion, we can also say that vascular brachytherapy has significant late loss in the proximal and more in the distal edges.” These differences, said Holmes, translated into improved lumen dimensions and better clinical outcomes with the drug-eluting stent (DES). He concluded that the SES was superior to VBT in reducing the primary end point of target vessel failure. Published in JAMA in 2006, the final conclusion stated that SES resulted in superior clinical and angiographic outcomes compared to VBT for the treatment of bare metal stent restenosis .

Later that same year, JAMA published results from TAXUS V ISR . The study demonstrated superiority in efficacy with the Taxus paclitaxel-eluting stent (PES) (Boston Scientific, Natick, MA) when compared to VBT using a beta emitter. As a result, physicians abandoned the use of VBT for ISR and DES became the default therapy.

In 2008, JACC: Cardiovascular Interventions published the 3-year follow-up data from SISR. At 3 years, the Cypher SES continued to improve target lesion- (TLR) and target vessel revascularization (TVR) survival rates when compared to those in patients treated with VBT. Stent thrombosis rates, however, were higher in the SES group (3.5% vs. 2.4%) . Five-year SISR follow-up data are currently published in the American Heart Journal . Surprisingly, at 5 years, clinical differences were no longer seen between SES and VBT, including survival free from TLR or from target vessel failure (TVF). There were no longer significant differences in the rates of definite/probable stent thrombosis; although rates increased with Cypher to 5.9% yet remained the same with VBT at 2.5% . If the rate of late stent thrombosis in the Cypher group continues to rise over the next couple of years, it could become statistically significant when compared to VBT. Given the large population that has been treated with SES for ISR, public concern regarding potential late events in this study population could arise.

Regarding the Taxus V ISR study, there is no available information on the event rate at 5 years. Changes in outcome and conclusions from year 1 to year 5 have been reported in the SIRTAX trial. SIRTAX was a randomized clinical trial in which 1012 patients were assigned treatment with either SES or PES. The primary clinical end point, the composite of cardiac death, myocardial infarction, and TLR at 9 months, was published in the New England Journal of Medicine in 2005, and showed a significantly lower major adverse cardiac event (MACE) rates with SES compared to PES, a difference largely driven by a significant reduction in TLR. Further, clinical benefit in favor of the SES was associated with significantly lower in-stent late loss compared to PES at 8 months. Patients were then followed clinically on an annual basis up to 5 years, and a second, repeat angiography was performed in willing patients .

Five years later, SIRTAX LATE reported that there are no longer differences between these two first-generation stents in terms of MACE. The angiographic result, which was superior in favor of SES at 1 year diminished over the 5 years with a greater increase in late loss with SES. In a landmark analysis comparing event rates up to 1 year, then between 1 and 5 years, Räber et al. found that regarding major adverse cardiac events (MACE), any real difference between the two stents vanished between 1 and 5 years, with the event curves actually crossing at around the two-year mark . Another principal finding was related to the risk for very late stent thrombosis, which was ongoing between 1 and 5 years, with a cumulative rate of definite stent thrombosis of 4.4% at 5 years.

Recent data from the head-to-head first-generation versus second-generation DES trials continue to support the dynamic change in the results across longer follow-up time periods.

For example, in the Endeavor IV trial comparing the Endeavor DES to Taxus DES, the 9- and 12-month results showed that the Endeavor DES was non-inferior to the Taxus DES, with no significant clinical differences despite higher in-stent late loss in the Endeavor DES, leading many to predict that this would lead to higher rates of revascularization over the long term. Interestingly, the 3-year outcomes indicate that the TVF rates show a strong but statistically non-significant trend favoring the Endeavor DES . Strikingly, the rates of death and myocardial infarction were significantly lower with the Endeavor DES; while stent thrombosis was higher in the Taxus group. In a landmark analysis, looking only at stent thrombosis between 360 and 1080 days, only 1 stent thrombosis occurred in an Endeavor-treated patient, compared with 11 in the Taxus group.

Similar differences were noted in the LEADERS trial where 1707 patients with 2472 lesions were randomly allocated to receive either a biodegradable polymer biolimus-eluting stent (BES) (857 patients, 1257 lesions) or a durable polymer SES (850 patients, 1215 lesions). At 4 years, biodegradable polymer BES was non-inferior to durable polymer SES for the primary end point. However, the landmark analysis shows that beyond 1 year and up to 4 years, there was a significant reduction in the risk of cardiac death or myocardial infarction with the BES .

Finally, we’ve watched very carefully the reports from PARTNER comparing surgical aortic valve replacement (SAVR) to transcatheter aortic valve replacement (TAVR) in patients with severe aortic stenosis. There were higher (borderline significant) stroke rates in the TAVR group compared to the SAVR group at 1 year. Such findings were associated with great controversy. At 2 years, with more strokes in the SAVR group, significant differences were no longer seen between groups .

While device approval for marketing can be derived from 1-year clinical data, it is imperative to redefine the long-term follow-up of the pivotal trials to at least 5 years since this longer follow-up may result in changes to indication and labeling for use. On the other hand, we don’t want to hold up approval for 5 years. This would result in different levels of approval and specific labeling that may be granted on a longitudinal axis of the follow-up period and will have to be built into the trial design. The question of ‘how long’ will always remain; yet for most of the emerging technologies, 1-year follow up will make it too early to call.