The FDA has a really important question and wants your advice.
This is not a fairy tale. This is a real-life story.
Hepatitis B is a serious disease. A company (Dynavax) has a new hepatitis vaccine that induces hepatitis antibodies more vigorously than existing vaccines and does so after 2 doses (instead of the usual 3). The vaccine works through a unique adjuvant. The serological advantages of the Dynavax vaccine were demonstrated in a randomized trial of >8000 patients; about 5600 people received the new vaccine and about 2800 people received the existing standard.
Why does the FDA need your help?
In the trial, an acute myocardial infarction occurred in 14 people in the Dynavax group, but in only one person receiving the conventional vaccine. The events were confirmed by adjudication. Since the Dynavax group was twice as large, the risk of acute myocardial infarction in the trial was seven times greater with the new vaccine. The FDA wants to know if the new vaccine should be approved for use in millions of people.
What do you say? What recommendation would you make?
If you think this is just hypothetical, think again. On July 28, 2017, the FDA convened a public advisory committee meeting to consider this exact question. The members of the committee consisted primarily of experts in infectious diseases and immunology. I was the only cardiologist on the committee.
If the 14:1 imbalance was due to the play of chance, then the issue of myocardial infarction risk was spurious, and the vaccine should be approved. However, if the 14:1 imbalance reflected a real increase in cardiovascular risk, then approval of Dynavax vaccine would be problematic.
Was it biologically plausible for the new vaccine to cause heart attacks?
The new adjuvant in the vaccine caused an inflammatory response (of uncertain duration), and inflammation is an important cause of rupture of atherosclerotic plaques. So a causal linkage was not out of the question.
Was the imbalance in myocardial infarctions due to the play of chance?
That was a good question, but it was impossible to know. Many might think that calculation of a P value would help, but it wouldn't. P values have a place in clinical trials, but not when the number of events is so small and the number of comparisons is so great. So no one asked for or showed any P values during the meeting. Everyone agreed that statistics could not resolve the uncertainty.
If you wanted to know if the 14:1 imbalance represented a real risk, you needed more information. You needed comparative data in 50,000 people. The fastest way of obtaining that evidence was through a post-marketing trial. But a post-marketing trial was possible only if the vaccine was approved for public use.
So what recommendation would you have made to the FDA?
The FDA asked the committee if there was reasonable evidence that the vaccine was safe. On July 28, the committee vote 12-1 (with 3 abstentions) in favor of the safety of the new vaccine. I was one of the three abstentions. Most of the committee believed that the vaccine's serological advantages outweighed the uncertainty, but the vote is non-binding. The FDA will decide on the new vaccine by August 10.
Why did I abstain? Based on the available data, it was impossible for anyone to know if the imbalance in myocardial infarctions was real or spurious. So although the question was fascinating and the discussion was terrific, my vote wasn't that complicated.
There is a simple rule in life: if you don't know, you should say that you don't know.
Disclosures
Packer has recently consulted for Amgen, Boehringer Ingelhim, Cardiorentis and Sanofi. He was one of the two co-principal investigators for the PARADIGM-HF trial (sacubitril/valsartan) and currently chairs the Executive Committee for the EMPEROR trial program (empagliflozin).