he Food and Drug Administration recently had a very important question about a new hepatitis vaccine and wanted advice.

Hepatitis B is a serious disease that can damage the liver. A company called Dynavax has developed a new vaccine that stimulates the body to produce antibodies against the hepatitis B virus more vigorously than existing vaccines do, and with just two doses instead of the usual three. The vaccine works through a unique adjuvant — a substance that improves the body’s immune response to the virus. The advantages of the Dynavax vaccine were demonstrated in a randomized trial of more than 8,000 patients: About 5,600 received the new vaccine and about 2,800 received the existing standard hepatitis B vaccine.

During the trial, though, 14 people in the Dynavax group had heart attacks compared to just one in the conventional vaccine group. Since the Dynavax group was twice as large, the heart attack risk was seven times higher with the new vaccine.


The FDA wanted to know if the new vaccine should be approved for use in millions of people.

To help make a decision, it convened a public advisory committee meeting on July 28. The members of the committee consisted primarily of experts in infectious diseases and immunology. I was the only cardiologist on the committee.

If the 7-1 imbalance was due to the play of chance, then the issue of myocardial infarction risk was spurious, and the vaccine should be approved. But if the 7-1 imbalance reflected a real increase in cardiovascular risk, then approving the Dynavax vaccine would be problematic.

The advisory committee needed to consider whether it was biologically plausible for the new vaccine to cause heart attacks.

The new adjuvant in the vaccine caused an inflammatory response of uncertain duration. We know that inflammation causes atherosclerotic plaques in coronary arteries to rupture — the event that triggers most heart attacks. So a causal link between the vaccine and heart attacks wasn’t out of the question.

Most importantly, we needed to decide if the imbalance in heart attacks between the two groups could have been due to the play of chance. That was a great question, but one that was impossible to answer. Many might think that calculating 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 so great. So no one asked for or showed any p-values during the meeting. Everyone agreed that statistics could not resolve the uncertainty.

To know if the 7 -1 heart attack imbalance represented a real risk, we’d need comparative data in 50,000 people. The fastest way of obtaining that evidence would be through a post-marketing trial. But a post-marketing trial would be possible only if the vaccine was approved for public use.

The FDA asked the committee if there was reasonable evidence that the vaccine was safe. Twelve committee members voted in favor of the safety of the new vaccine, one voted against it, and three abstained. I was one of those who abstained. The vote is nonbinding. The FDA had originally said it would decide on the new vaccine by Aug. 10, but late yesterday asked Dynavax for more information about the company’s post-marketing study. An FDA decision about the vaccine is now expected in November.

Why did I abstain? Based on the available data, it was impossible for anyone to know if the increase in heart attack risk in the Dynavax group was real or spurious. So although the questions were fascinating and the discussions terrific, my vote wasn’t that complicated.

There is a simple rule in life: if you don’t know, you should say you don’t know.

Milton Packer, M.D., is the Distinguished Scholar in Cardiovascular Science at Baylor University Medical Center at Dallas. A version of this article originally appeared as part of his “Revolution and Revelation” column on MedPageToday.

This article was updated to reflect the FDA’s request for more information about the Dynavax vaccine.

Leave a Comment

Please enter your name.
Please enter a comment.

  • I don’t understand why statistics cannot help to resolve the uncertainty. There must be data out there about the base probability of getting a heart attack in a similar population. But in any case, you have a large control, so it is very likely that the probability should be around 1/2,800. Given that probability, the odds of observing more than 13 cases in a population of 5,600 is less than 1 out 300 million. You can go Bayesian about it and put a prior on the probability (for example a Beta(2, 2801)), but you will get to the same result: it is very very certain that the new vaccine increases the risk of a heart attack.

    From a societal perspective, how much of an increase can you tolerate depends how serious a Hep B infection is compared to a heart attack, as well as how rare Hep B infections are and how effective is the vaccine compared to the old vaccine. But in any case it seems like all this could actually be quantified.

    From a personal perspective, with an old vaccine already at around 95% effectiveness, the extra trip to the doctor and extra shot are totally worth avoiding the risk of a heart attack.

    • Jorge-

      Two factual errors in your response:

      The annual risk of a heart attack in the general population is known and is much higher in the general population than it was in the control group. It ranges from about 1/500 in parts of the Northwest US to as high as 1/125 in parts of the South (above age 35 – source: https://www.cdc.gov/heartdisease/images/heart_disease_deaths.png). The Dynavax vaccine was pretty close to midrange for this. The issues is that the control group was significantly LESS frequent than normal. I guess you could argue that Engerix is protective against MI and Heplisav is not, but this is one of the reasons the 12 thought the difference was an anomaly.

      Second, the current vaccines do not have anywhere near a 95% efficacy rate – especially if you take into account compliance to a 3-dose schedule. Heplisav is CLEARLY superior in terms of both compliance and efficacy. The data speaks for itself there.

      Ken Pittman, MD

    • The article was edited for STAT’s general audience with Dr. Packer’s help and his final sign-off.

      Patrick Skerrett
      Editor, First Opinion

  • this is a refreshing opinion piece by dr. packer. if there is already a treatment available, then there should be no rush to approve a vaccine that is so much more dangerous. i find it interesting that he states “if the 7-1 imbalance reflected a real increase in cardiovascular risk, then approving the dynavax vaccine would be problematic.” does that mean that even if the increase in the risk of heart attack was found to be real and not spurious, the vaccine could still be approved? although reasonable, objective minds could differ, i truly hope that the fda required members of the public advisory committee to openly declare any potential conflicts of interest. the fact that 12 committee members voted in favor of this is concerning.

    • Conflicts of interest were posted and no material conflicts of interested were included for any member of the committee. I do not recall seeing Dr. Packer’s work with GlaxoSmithKline (makers of Engerix) in the past listed, but disclosure of this may not have been required.

  • To DRJRH:
    The efficacy data for Dynavax’s Hep B vaccine is much better than either one on the market. It is also a 2 dose vaccine rather than a 3 dose one like the ones on the market. Therefore compliance with it is MUCH better. Only ONE out of 3 studies showed the flag that Dr. Packer is talking about and there were other very plausible explanations related to grouping characteristics. I have no issues with Dr. Packer’s abstention vote – Dr. Packer is a very smart man and he is correct that we can’t be certain until a postmarketing study. However – he did leave out the facts that Hepatitis B is a devastating illness in itself currently impacting 257 million people in the world. He left out the fact that this is DESPITE having two vaccines on the market already. He left out the fact that many world and community health experts testified about the need for this vaccine at the ADCOM he was a part of – and most of them are perfectly familiar with the risk of uncertainty. The opinion of the 12 people that voted in favor of it (despite Dr. Packer being VERY vocal about the uncertainty) was that it was much more likely to be a “spurious” finding (this word was used many, many times by panel members). Yes, these people weren’t cardiologists. However, they WERE experts in immunology, communicable diseases, etc. Dr. Packer is a very intelligent man, but every single person on the panel was of near equal intelligence. Intelligent people don’t always come to the same conclusion. 12 out of 16 came to the conclusion that the vaccine has shown sufficient safety data for approval. 3 people (including Dr. Packer) shared his opinion of “I don’t know.” 1 person believed that the vaccine was not sufficiently shown to be safe. If the same approach was taken by Dr. Packer and the others that voted to abstain and no, they would have done the same with previous Hep B vaccines as well. I say this because those vaccines also had areas in their pre-approval studies that were uncertain of safety. In fact, I’d argue that EVERY vaccine and drug approval carries some degree of uncertainty. If we took Dr. Packer’s approach to abstain when there was any degree of uncertainty, then perhaps all new drugs and vaccines would get “I don’t know” votes. Again, I respect Dr. Packer’s vote, but there are certainly excellent arguments for those that voted “Yes” as well and they aren’t taking to blogs to describe their reasoning.
    Ken Pittman, MD
    Pediatrician and Child Psychiatrist

    • I see you’re a psychiatrist. Which means your expertise is not in cardiology, neurology, or Immunology. I did notice that you went on a pro any vaccine advertising campaign in your comment. That’s inappropriate for this venue

    • I also noticed your pediatrician supposedly. I would be curious to know, since you did go on a pro any vaccine advertising campaign, how much of a bonus you get per fully immunized patient ,according to the newest schedule that comes out every year. I’m very curious to know that. Don’t tell me that I’m ignorant and uneducated, I’ve been studying and following this and the things the numerous ingredients cause for years. Rest assured, if I don’t understand something I’ll stop and I’ll study one particular thing until I fully understand it.

    • Lisa-

      To answer your comments/questions:
      My experience as a Pediatrician and Child Psychiatrist is certainly relevant. Pediatrics is where the majority of vaccines are given (although this particular vaccine is only going to be indicated for adults at this time). Mental illness is associated with communicable diseases including Hepatitis B. So, yes, both are relevant and appropriate for this venue (which is a website for health and medicine that is discussing the approval of a vaccine).

      Regarding my “Pro Any Vaccine Advertising Campaign”
      I’d like you to comment on what part you felt was that. My comment was that no drug or vaccine would ever be approved if we only voted “Yes” for safety on things that had no uncertainties. That is not advertising anything – it’s a simple fact that uncertainty will always exist.

      As for my actual use of vaccines – I have not administered vaccines to my patients other than the flu vaccine in about 6 years. This is because I have been primarily practicing Child Psychiatry over that time. I would not hesitate to give vaccines by the recommended schedules, though – except in very rare instances where there is a strong family history of autoimmune issues.

      I did neglect to disclose that I am a small investor in Dynavax. That was an oversight on my part as I came to this page from a page where my investment in Dynavax is well disclosed. For full disclosure, Dr. Packer did not disclose in the article that he has received money from GlaxoSmithKline (who made Engerix-B, the primary competitor for this vaccine). This information is also disclosed in other forums. Neither my disclosure nor his are necessary in this forum, but I believe knowing both are appropriate.

      Ken Pittman, MD
      Pediatrician and Child Psychiatrist

    • Now that’s an interesting comment about family history of autoimmune disorders.

      I wonder if others have the same idea.

      I would say though this heuristic can be misleading when the family is originally from another country, both because of a potential lack of sufficient medical records/low quality of medical documentation, and because America has a very high rate of autoimmune disease not found in many other countries — something that may be environmentally triggered, instead of genetically.

  • more test ,and this is why drugs are so high in the USA , hey cars and bees will kill more people than this drug -if it does , healthy people drop dead of heart-attacks , approve the drug and stop costing people money , or a delay so the friends and family can get in cheap , , funny senators and congressmen make 175,000 , a yr but they all are millionaires , how , cause of delays like this , WINK,WINK, NOD,NOD ,

  • Given that there are effective alternative hepatitis B vaccines available in the marketplace and that you could not rule out a seven-fold increased risk for a potentially life-threatening complication for a product to be administered to generally healthy subjects, then the correct answer should have been that the product should not be approved for safety reasons.

    • IMHO (as a PhD candidate in Biostatistics and Epidemiology) I do not agree at all.
      First of all, in spite of the chance to use a Bayesian approach, I would not say that 1 (or seven) cases could be taken as a strong evidence supporting an increased risk – especially if, as someone else noted, the control group of the RCT had a lower-than-expected rate of CVDs: it could be pinpointing some sort of selection bias, rather than a RR>1.

      Moreover, no perfect drug exists, and an authority MUST also take into account the impact when evaluating marketability of a new drug. In other words, a (in any case) small risk of CVDs, even if higher than already commercialized HepB vaccines, could be balanced by benefits – which is something that usually comes out clearly when you look at a population level and not at individual risks. For instance, a roughly 0.1% estimated risk of acute CVDs on two doses could be “easier to care about” than managing less heavy ADRs on the third dose which is prescribed with the GSK’s vaccine on the market.

      Disclosure: I definitely have no conflict of interests to disclose. I have never worked nor being beneficiary of a single euro from pharma companies. It is just my opinion about what evidence and epi logic suggest on this case.

Sign up for our Morning Rounds newsletter

Your daily dose of news in health and medicine