December 3, 2025
I am Dr Diana Zuckerman, president of the National Center for Health Research. Our nonprofit public health research center focuses on the safety and effectiveness of medical products and we do NOT accept funding from any entities with a financial interest in our work. So the Center has no conflicts of interest, but my father worked for J & J his entire career and I inherited J & J stock from him. However, J & J won’t be happy with my testimony today.
My expertise is as a former faculty member and researcher at Yale and Harvard and as an expert in FDA policy in the U.S. Congress, the White House, and at several nonprofit organizations.
As a Breakthrough device, the FDA gave priority review to the V-Wave Ventura shunt, because it is a novel technology with the potential to provide more effective treatment for certain patients. I’m sure you noticed the 2 pages of exclusion criteria to participate in this study; this device was studied on a relatively small proportion of ambulatory Class IV heart failure patients with reduced or preserved left ventricular systolic function.
The FDA points out that breakthrough devices must meet the same standards as other medical devices to get on the market: a reasonable assurance of safety and effectiveness. FDA may accept greater uncertainty of the benefit-risk profile if “appropriate”, such as a device that treats a life-threatening disease when no alternative treatments are available.
However, in this case, the device was not more effective than a sham control in the overall study – in other words, not more effective than placebo.
Should a device that is no better than placebo be approved, even if its short-term safety is no worse than placebo? That makes no sense, because it certainly adds to the cost of medical treatment, and when used by physicians who are less experienced than those participating in this clinical trial, one has to assume the risks could increase as well
And, if it is approved it could legally be used for any heart failure patient. In fact, device companies have been pressuring Congress to REQUIRE Medicare to pay for any breakthrough devices on the market. For that reason, your work is especially important today.
I am here to discuss the scientific evidence and the statistical issues, because that is my expertise. But I have a personal stake as well, because both my parents and all 4 of my grandparents died of cardiovascular disease and my mother lived for years with heart failure. So, the fact that control patients sometimes did better than shunt patients concerns me.
I agree that the study design is very impressive: a randomized blinded sham-controlled trial.
The Primary Effectiveness Endpoint was a composite including All-cause death • Cardiac transplantation or LVAD implantation • Heart failure hospitalization of ≥6 hours• 5 pt Change in Kansas City Cardiomyopathy Questionnaire (KCCQ) score, and other variables
Secondary Endpoints were to be tested ONLY if the primary endpoint was statistically significant.
However, “the primary effectiveness endpoint was not met and there was no signal of Shunt benefit in the primary effectiveness endpoint results. Rates for the composite endpoint components of death, cardiac transplantation/LVAD, HF Hospitalization, and worsening outpatient HF events were generally similar between the Shunt and Control groups, and at all timepoints through 2 years, changes in KCCQ (Quality of life) scores were similar between the Shunt and Control groups.
All-cause death and HFH rates numerically favored the Control group, while cardiac transplantation/LVAD and worsening outpatient HF event rates favored the Shunt group, but those differences were small.
The calculations of the individual components and the method of calculating composite rates were not pre-specified. That’s why we can’t draw conclusions regarding statistical significance.
The company did its best to figure out who was most likely to benefit from the shunt. Those analyses are important for the company to help them either improve the shunt or specify an indication that shows a benefit compared to the control group. However, as the FDA pointed out, those post-hoc analyses should be considered exploratory or hypothesis generating. They should not be used to justify FDA approval given the failure to meet the primary endpoint.
We all know that post-hoc analyses have many shortcomings, especially Type 1 errors and lack of control for multiple comparisons, and when they are conducted on subgroups with inconsistent results, it is impossible to draw conclusions. As FDA pointed out, even 4 patients in the Control group can change the results because of the smaller subgroups being analyzed. This shunt should be studied in larger groups of specific types of patients BEFORE it can be considered for approval.
What about the Win Ratio data? As noted in articles published last year in Circulation and elsewhere, the Win Ratio has important limitations. #1: It is unclear how to prioritize component events. Is myocardial infarction worse than stroke? Is 1 early hospitalization worse than 2 later hospitalizations? #2: combining short-term symptoms with major outcomes later in the trial in a single metric may conflate short- and long-term efficacy. #3: Including additional end points may increase statistical power, but the result may be driven by unimportant events with small differences.
In conclusion: For all components of the composite endpoint, the sponsor expected that event rates would favor the Shunt group so it was notable that it didn’t. Most important: The similarity in CV death rates between treatment groups absolutely does not support a mortality benefit associated with Shunt use.
Thank you for the opportunity to share my views.