FDA Approves Optimizer Smart Device for Class III Heart Failure Patients
Drug/Device: OPTIMIZER SMART device
Sponsor: Impulse Dynamics (USA), Inc.
Committee: Medical Devices Advisory Committee
Indications：to provide cardiac contractility modulation (CCM) for class III heart failure patients who are not responding to optimal medical therapy.
Question 1: Is there reasonable assurance that the OPTIMIZER System is SAFE for use in patients who meet the criteria specified in the proposed indication?
Vote Results Yes 12 No 1 Abstain 0
Question 2: Is there reasonable assurance that the OPTIMIZER System is EFFECTIVE for use in patients who meet the criteria specified in the proposed indication?
Vote Results Yes 11 No 2 Abstain 0
Question 3 ：Do the BENEFITS of the OPTIMIZER System outweigh the RISKS for use in patients who meet the criteria specified in the proposed indication?
Vote Results Yes 12 No 0 Abstain 1
Related Documents:
Background:
The Pre-Market Application for the OPTIMIZER SMART device was submitted with two studies.: FIX-HF-5 and FIX-HF-5C. The FIX-HF-5 study did not demonstrate a statistically significant difference in the primary effectiveness endpoint which was an improvement in VAT on CPX testing. However, it did meet its primary safety endpoint and did not raise new safety concerns. A post-hoc subgroup analysis on the subjects with an LVEF >25% and NYHA class III symptoms showed an improvement in VAT, as well as pVO2, NYHA class and quality of life which generated the hypothesis for the FIX-HF-5C study. FIX-HF-5C was designed as a confirmatory study. The goal of FIX-HF-5C was to confirm the subgroup analysis findings of the FIX-HF-5 study that patients with an LVEF of 25-45% and NYHA class III/IV heart failure symptoms benefit from the delivery of CCM therapy.
Unique Challenges for Optimizer:
Bayesian methodology helped to reduce the sample size by 30% in the confirmatory study and borrow strong results from the previous study. FDA endorsed the Bayesian analysis.
Initially, the study planned to randomize 230 subjects. Due to the breakthrough Device designation, FDA agreed with the sponsor’s request to reduction in sample size based on implementation of a Bayesian analysis. The sample size reduced to 160 subjects, plus borrowing 69 subjects from 229 subjects of post-hoc identified subgroup of interested (significant treatment effect on the endpoint of peak VO2) from the previous study FIX-HF-5 at 30% weight (229 with 30% weight, 229x30%=69). Pre-specified 30% weight ensures prospective FIX-HF-5C data not dominated by prior FIX-HF-5 data. Though 11% weight had provided statistical significant result, the 30% borrowing figure was arrived at after multiple simulations for acceptable results in multiple realistic clinical trial scenarios. The sponsor selected method and weight through multiple simulation scenarios in collaboration with FDA and an external statistics expert group.
Bayesian analysis with borrowing showed statistically significant result for primary endpoint, while analysis without borrowing was not significant. FDA accepted the pre-planned Bayesian analysis result.
Bayesian Mixed effect model with borrowing was used for calculation of Primary Effectiveness Endpoint. The subgroup from FIX-HF-5 has the similar patient demographics with FIX-HF-5C. It is considered being statistically valid if analysis incorporates relevant and prior information from previous study. In Bayesian model, the primary effectiveness endpoint, the mean Peak VO2 difference, was 0.84 (0.12, 1.55) with 0.989 Posterior probability Benefit(>0.975). While with FIX-HF-5C alone, Peak VO2 treatment effect was not statistically significant.
All secondary effectiveness endpoints and safety analyses conducted on FIX-HF-5C data alone using frequentist methods with 1-side lower 95% CI. Improvements in quality of life and NYHA class were observed with CCM therapy. Interestingly, for quality of life and NYHA class, the treatment effects using FIX-HF-5C data alone (without borrowing) are superior to both FIX-HF-5C data with borrowing and pooled data (FIX-HF-5 and FIX-HF-5C), though the former 95%CI is bigger than the latter ones.
The primary safety endpoint, 89.7% CCM-treated patient complication-Free rate met the performance goal which was pre-set up as greater than70%. The results for all pre-specified secondary safety endpoints demonstrated no significant safety concerns with all p-values >0.025. Overall survival, freedom from CV death and severity of AE, were similar between treatment arms.