• Svelte DES Systems Fail to Meet Prespecified Noninferiority Criteria vs. Established DES

    Study's lead investigator makes a case that the novel stents really are comparable to Xience, Promus

    Two ultra-low-profile drug-eluting stents (DES) fell short of demonstrating the prespecified noninferiority criteria in comparison with more-established, commercially approved DES, even though “the totality of the evidence” shows that these new stents are, in fact, noninferior, according to the study’s lead investigator.

    Dean J. Kereiakes, MD, of The Christ Hospital Heart and Vascular Center, Cincinnati, presented the results of the OPTIMIZE trial on Saturday at the TCT Connect virtual conference.

    The ultra-low-profile fixed-wire Slender Integrated Delivery System (IDS) and rapid-delivery Direct RX DES systems, both manufactured by Svelte Medical Systems, are designed to facilitate transradial access and direct stenting, Kereiakes and co-authors wrote in an abstract presenting the OPTIMIZE study.

    OPTIMIZE is a prospective, single-blind, randomized, international investigational device exemption trial comparing the safety and efficacy of the Slender IDS and RX with that of the Xience (Medtronic) or Promus (Boston Scientific) DES in patients with ischemic heart disease and no more than three de novo stenotic lesions that are no more than 34 mm long in no more than two native coronary arteries with reference vessel diameter of 2.25 to 4 mm that were amenable to percutaneous coronary intervention (PCI).

    A total of 1,630 patients were randomized to receive the Svelte or control DES. The primary endpoint was to demonstrate noninferiority of the Svelte stent systems with regard to target lesion failure (TLF), a composite of cardiac death, target vessel myocardial infarction (TVMI, including both Q-wave and non-Q-wave) and clinically driven target lesion revascularization (TLR) at 12 months. The TVMI component of the endpoint also included periprocedural MI, which was defined as creatine kinase-MB fraction (CK-MB) or troponin more than 3 times the upper limit of normal within 48 hours of the procedure.

    Secondary endpoints included components of TLF; target vessel failure; major adverse cardiac events; Academic Research Consortium-defined stent thrombosis; and lesion, device, procedure and direct-stent-strategy success.

    The expected TLF rate based on the EVOLVE II trial was 6.5%. Noninferiority margin was set at 3.58%, and a one-sided alpha p-value <0.025 in the intention-to-treat (ITT) analysis would mean that the Svelte stents were noninferior to the control stents.

    Baseline characteristics between the control and Svelte DES groups were well-matched (mean age, control 65.8 years vs. Svelte 65.1 years; men, 70.8% vs. 72.7%; Caucasian 82.4% vs. 81.4%; Asian 11% vs. 10.9%; smoking history, 61.3% vs. 63.7%; diabetes, 30.7% vs. 28.5%).

    The same was mostly true of procedural characteristics, except that more patients in the Svelte group had three lesions treated (3.5% vs. control 1.6%), and the treated lesions were longer in the Svelte group (14.88±7.04 mm vs. control 14.25±7.52 mm).

    The Svelte stents failed to demonstrate noninferiority to the control stents with regard to the primary endpoint, TLF at 12 months (Svelte 10.3% vs. control 9.5%; difference +0.8%; upper bound of 95% confidence interval, 3.8%, noninferiority margin, 3.58%; p-noninferiority = 0.034; prespecified p for noninferiority = 0.025).

    With regard to the TLF components at 12 months, the Svelte and control stents showed no significant differences.

    Kereiakes noted that the TLF rate was driven by the TVMI rate (control stents 8.22% vs. Svelte stents 9.31%; p=0.48). Of the TVMIs in both arms, 90% were periprocedural MIs. He added that 25% of patients with troponin assays accounted for 80% of TVMIs. Of the troponin-positive patients, 3.8% had electrocardiogram changes and 87.5% were discharged without delay, he said.

    The overall study’s TVMI rate of 8.8% was “unprecedented,” Kereiakes said, reflecting the frequency of troponin use as a biomarker, which he said contributed to the study’s high TLF rate (overall 9.9% vs. an expected 6.5% rate) and contributed to “effectively underpowering the OPTIMIZE study.”

    These findings prompted the investigators to analyze relative risk and assessment in OPTIMIZE in comparison with other investigational device exemption studies. An independent post hoc analysis was also conducted to determine whether the OPTIMIZE relative risk was within the prespecified protocol noninferiority margin of 1.55. This margin was assigned based on the ratio of noninferiority margin with estimated TLF.

    The analysis found that the relative risk was 1.09 (95% confidence interval, 0.81-1.46), meaning the upper bound of the confidence interval was within the prespecified margin on the post hoc analysis.

    Based on this finding, Kereiakes said, the Svelte stents were noninferior to the control stents (p=0.009). Kereiakes speculated that future studies will probably not use the same protocols as OPTIMIZE because of how this study turned out.

    Some panelists at the press conference announcing the results agreed with Kereiakes’ points.

    Robert W. Yeh, MD, MBA, of Beth Israel Deaconess Medical Center, Boston, who was not involved in the study, said he thinks that the U.S. Food and Drug Administration (FDA) would consider not just the primary endpoint but the totality of the evidence when it determines whether to approve the Svelte stents for marketing in the U.S. He added that he has agreed to serve on the FDA’s Circulatory System Devices Panel, which will provide him with further insight.

    Still, the study failed to demonstrate noninferiority of the Svelte stents based on the prespecified endpoints.

    Robert A. Byrne, MB, BCh, PhD, of Mater Private Hospital, RCSI University of Medicine and Health Sciences, Dublin, who also was not involved in the study, underscored that point.

    “At the end of the day, you have to look at the primary endpoint, and there’s reasons behind it, but the criteria weren’t met,” he said. “In our study (COBRA-REDUCE) – I didn’t show it – but if we had chosen cardiac death instead of total death, we would have met noninferiority. But you have to base your primary interpretation on the primary endpoint, and granted, the regulatory authorities then have to take a look afterwards and see, are they going to look at the totality of the evidence?”

    Svelte Medical Systems provided all funding for this trial.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Review our Privacy Policy for more details