Author + information
- Published online January 2, 2017.
- Ron Waksman, MD∗ ()
- ↵∗Reprint requests and correspondence:
Dr. Ron Waksman, MedStar Washington Hospital Center, Division of Interventional Cardiology, 110 Irving Street, NW, Suite 4B-1, Washington, DC 20010.
Diabetes affects more than 29 million people in the United States. The number of diabetic patients presenting with coronary artery disease is on the rise, accounting for more than 30% of patients referred for revascularization (1). The treatment of diabetic patients presenting with coronary artery disease remains challenging. Randomized clinical trials, such as the FREEDOM (Comparison of Two Treatments for Multivessel Coronary Artery Disease in Individuals With Diabetes) trial (2), have demonstrated the superiority of coronary artery bypass grafting compared with percutaneous coronary intervention, especially in patients with multivessel disease or high SYNTAX score. Nevertheless, advances in stent technology, especially with drug-eluting stents (DES) and most recently with the newer generation of DES, have made stents a viable and less invasive alternative therapy to coronary artery bypass grafting for patients with less complex anatomy.
The initial experience with first-generation DES, such as the sirolimus-eluting stent (SES) and the paclitaxel-eluting stent (PES), was promising for diabetic patients, especially when compared with bare-metal stents (3). The sponsors of the PES claimed superiority over SES for diabetic patients, but this was based only on observational studies (4). The Food and Drug Administration (FDA) at that time was not convinced that these data were sufficient to support an indication on labeling.
Shortly after the introduction of first-generation DES for clinical use, the enthusiasm for them was tempered by the rise in stent thrombosis and late recurrences, especially for patients with insulin-dependent diabetes mellitus (DM) (3). Those findings motivated scientists and bioengineers to try to develop a stent specifically for diabetic patients, which led to second- and third-generation DES. Although the newer DES have been shown to deliver numerically better outcomes across most subgroups of patients and lesions, including diabetic patients, this has not risen to the level of significant superiority for the diabetic population.
As part of the development of second-generation DES, the FDA became interested in additional outcomes data in diabetic patients, and sponsors sought a specific indication for diabetic patients on labeling. The sponsors and FDA worked collaboratively to develop a performance goal (PG) to address safety and effectiveness concerns associated with DES use in diabetic patients. The PG was based on pooled outcome data from first-generation DES in the diabetic population. Specifically, 1-year target-vessel-failure (TVF) outcomes for DM patients implanted with the Resolute Integrity zotarolimus-eluting stent (ZES) (Medtronic, Santa Rosa, California) were compared against a PG derived from a meta-analysis that included 6 trials with Cypher SES (Cordis Corporation, Warren, New Jersey) and Taxus PES (Boston Scientific, Natick, Massachusetts) along with pooled data for the Endeavor ZES. The rate of TVF—a composite endpoint that included cardiac death, myocardial infarction, and target vessel revascularization—for the diabetic ZES patients was 7.8% (upper 95% confidence interval 9.51%). That met the PG of 14.5% (p = 0.001), and Resolute was granted the new indication for DM.
A potential impact of this labeling was increased usage of the Resolute ZES stent in diabetic patients, as some physicians now erroneously believed that it had been shown to be superior to other second-generation DES for the diabetic population.
It also led to requests from other manufacturers of second-generation DES that the FDA grant their products an indication for use in diabetic patients based on similar performance goal criteria as was used for the Resolute DES. Abbott conducted a prospective analysis designed to evaluate the safety and effectiveness of the XIENCE family of stents (Abbott Vascular, Santa Clara, California) for patients with DM using 1-year data from 4 clinical trials and 2 external registry databases. The TVF rate at 12 months was tested against a pre-specified PG of 14.8% (expected rate 8.6% plus a delta of 6.2%). The PG was met and the XIENCE stent was granted the desired labeling. Following that, Boston Scientific jumped into the labeling race and received an indication for diabetic patients for both of their DES programs, Promus and Synergy. Specifically, Promus Premier was granted an indication for DM patients on the basis of a post-approval study with a PG of 12.6% and a delta of 4.2%. In the case of Synergy, the indication was granted on the basis of a substudy called EVOLVE II DM that pooled diabetic patient data from the EVOLVE II (EVOLVE II Clinical Trial To Assess the SYNERGY Stent System for the Treatment of Atherosclerotic Lesion[s]) randomized study with data from nonrandomized studies. EVOLVE II DM showed an overall target lesion failure (TLF) rate of 7.5% in the intent-to-treat subjects compared with a PG of 14.5% that was based on data from historic controls in randomized trials that involved everolimus-eluting stents.
The outcome of the labeling race is that all second-generation DES approved in the United States now carry an indication for use in DM patients. Therefore, the indication no longer carries a marketing advantage, and we still lack a robust, randomized clinical trial that supports the use of these stents in the diabetic population. Although the FDA may feel obliged to grant a diabetic indication to any DES stent that meets the PG due to the precedent established with Resolute, the question remains: Does it benefit diabetic patients? Answering this question is particularly important now, considering that the diabetes indication can mislead physicians to believe that the stent has been shown to perform better than its competitors for this population when the reality is that the PG is more of a reflection of overall improved performance compared with first-generation stents and not necessarily an indication of how it performs in diabetic patients.
With the recent approval of the Absorb scaffold, the need for a clearly defined method of determining whether DES and scaffolds benefit diabetic patients is even more pressing. The sponsor agreed with the FDA on a pre-specified, powered analysis to meet the PG. Data from diabetic patients who received ≥1 Absorb in the ABSORB II, III, and JAPAN randomized clinical trials and the single arm ABSORB EXTEND registry were pooled. The primary endpoint was TLF at 1 year following Absorb implantation compared with a PG of 12.7% and a delta of 4.5% noninferiority margin. In this issue of JACC: Cardiovascular Interventions, Kereiakes et al. (5) report a 1-year TLF rate in their pool analysis of 8.3%, which meets the pre-specified PG. The question is whether those findings support expanding Absorb’s indication on labeling to specifically include diabetic patients when the current labeling as written does not preclude use of the stent in those patients?
The report from Kereiakes et al. (5) provides alarming scaffold thrombotic rates of 6.5% and overall TLF rates of 15.5% in diabetic patients with reference vessel diameter <2.25 mm who received the Absorb scaffold. What the paper does not tell us are the TLF and stent thrombosis rates in the XIENCE arm of the randomized trials. At the FDA’s Circulatory Systems Device Panel meeting regarding Absorb in March, FDA reviewers presented rates of scaffold thrombosis in diabetic patients with a reference vessel diameter <2.25 mm from the ABSORB III study of 10.6% in the Absorb group compared with 4.4% in the XIENCE arm. The TLF rate was 23.9% in the Absorb group versus 15.6% in the XIENCE group. Interestingly, for diabetic patients with a reference vessel diameter >2.25 mm, there was no difference in the TLF rate between the Absorb and XIENCE groups, but the Absorb group had a higher thrombosis rate of 1.3% versus 0.6% in the XIENCE group (6).
Another reason PGs are problematic for determining performance of DES or scaffolds in diabetic patients is the heterogeneity of the disease. Different rates of stent thrombosis and TLF have been reported in patients with insulin-dependent versus non-insulin diabetes. Other confounders of diabetes, HbA1c levels, and whether the disease is diffuse or focal also may impact outcome and are hard to match when using PG.
The available DES technology is not specifically designed to address lesions in diabetic patients. Granting the current DES, including Absorb, an indication for diabetic patients based on PGs is not supported by science and can be potentially misinterpreted, especially considering that the second-generation DES perform similarly to each other in randomized clinical trials. The FDA should revisit the reliance on PG to support labeling indications, and instead encourage sponsors to address public health care issues related to approved second-generation DES as a class effect when they seek a specific indication on labeling as the agency did very successfully when it was able to convince sponsors to conduct the DAPT (Dual Antiplatelet Therapy) study (7). It is imperative that expanded indications on labeling of DES are scientifically valid. Industry, the FDA, and investigators should reform the current process for indication expansion.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Waksman is a consultant for Abbott Vascular, Biosensors International, Biotronik, Boston Scientific, Medtronic, and St. Jude Medical; is on the speakers bureau of AstraZeneca, Boston Scientific, and Merck; and has received grant support from AstraZeneca, Biosensors International, Biotronik, Boston Scientific, and Edwards Lifesciences.
- American College of Cardiology Foundation
- Centers for Disease Control and Prevention
- Ritsinger V.,
- Saleh N.,
- Lagerqvist B.,
- Norhammar A.
- Kereiakes D.J.,
- Ellis S.G.,
- Kimura T.,
- et al.
- Steinvil A.,
- Rogers T.,
- Torguson R.,
- Waksman R.