When Sponsors Stop Trials Everyone Loses
The ANTHEM HFrEF trial of vagal nerve stimulation was proceeding well when the sponsor stopped the trial. The story...and some thoughts about the research enterprise
Vagal nerve stimulation holds some promise to help patients with heart failure due to a reduced ejection fraction (HFrEF). The idea is that these patients have an overabundance of sympathetic stimulation, which is detrimental in many ways to the heart. Previous trials of VNS have been inconclusive.
A company called LivaNova has a novel VNS system. It looks a bit like a pacemaker with a lead to the carotid area for vagal nerve stimulation.
A pilot study was encouraging. FDA gave the company a breakthrough device designation and allowed an adaptive statistical design to minimize the number of patients in the trial. The idea behind this special designation is to speed the availability of innovative products.
JACC published the ANTHEM trial results as well as the trial’s story.
Trialists designed ANTHEM like many other HFrEF trials: a device arm vs a non-device arm (Note: no sham control). The primary endpoint was standard—CV death and hospitalization for HF.
The planned enrollment was 1000 patients. The trial had a data safety monitoring board that looked at interim results.
After enrolling 532 patients the sponsor—not the DSMB—stopped enrollment.
The primary efficacy endpoint occurred in 34.6% in the treatment group and 40.2% in the control group. The HR was 0.84; 95% CI: 0.62-1.12; P = 0.115), which did not meet statistical significance. The primary safety endpoint achieved 96.7% freedom from procedure- or device-related serious adverse events.
The device seemed to improve vagal effects: 24-hour ECGs showed increased heart rate variability.
The authors modified the statistical analysis plan and added a hierarchical analysis using a win-ratio which was positive for VNS but inconclusive. It was inconclusive because a) it was not prespecified and b) included a quality of life assessment and there was no blinding.
The authors noted that the trial’s DSMB evaluated results after 500 patients (as planned) and did not find reason to terminate the trial for futility. Two editorialists added this comment (italics mine)
At this point, the trial’s DSMB recommended proceeding with further enrollment because prespecified conditions for early FDA filing or futility had not been met.
A few months later, the company, however, did their own analysis and stopped the trial. At the trial’s presentation, lead author of ANTHEM, Marvin Konstam, MD described the decision this way:
“not associated with either study futility, safety concerns, or device malfunction,” it was financial.”
Comments
I may be wrong but my read of this was that FDA gave the company a good shot at approval. The device did pretty well, but not good enough for early filing. The trialists did not find futility and called for continuing the trial. The company weighed the costs of the trial against the probability of profit and made a financial decision to stop.
It does not seem like any rule was broken, but it’s bad for all involved; bad for patients with HFrEF who may benefit from VNS, super bad for the patients enrolled in the trial, and ugly for the research enterprise.
The story highlights the tension from having so much of our progress in cardiology dependent on profit from industry. This is not a statement against profit; clearly, profit motive has brought us many advances in modern medicine. I would not want to go back to the 1970s. But profit motive is not exactly congruent with scientific inquiry.
Drs. Dhruva and Kadakia writing in an editorial note that the company’s actions undermined the justification behind the FDA’s upfront investment of regulatory resources and risks. They suggested that FDA attach explicit conditions to breakthrough designations in the pivotal trial setting. Namely, that there be prospective agreement on permissible grounds for sponsor-initiated termination.
That sounds like a good tweak. But perhaps a better tweak would be that there be less phase 3 (regulatory) trials conducted by parties who stand to gain from positive results of the trial. How to accomplish that is not an easy task.
Yet until we rework the evidence-generation system, I suspect we will continue to see such things.
My Goodness. The primary efficacy endpoint occurred in 34.6% in the treatment group and 40.2% in the control group. And adverse events were 3%. The company was prudent not to continue. We don't need more marginal devices or meds. We should concentrate on the root causes of these diseases.
Another aspect of a decision like this is that it distorts the scientific record by omitting potential negative results. Researchers in general are more likely to publish positive than negative results - for reasons including but not limited to financial incentives. The result is that future meta-analysis or other statistical evaluations are skewed towards effectiveness.
Thus even if "no rule was broken" here arguably the early termination was still unethical.