Cardiology Topics to Watch for in 2019

Interviewer: Robert A. Harrington, MD; Interviewee: C. Michael Gibson, MD


February 13, 2019

Editor's Note:
Owing to technical difficulties, the audio file is not available for this episode of The Bob Harrington Show. Below is an edited transcript of the discussion. We apologize for any inconvenience.

Robert A. Harrington, MD: Hi. This is Bob Harrington from Stanford University on and Medscape Cardiology. At the end of 2018, Mike Gibson and I had the opportunity to talk about the trials and stories of 2018 in cardiology that caught our attention. We will now concentrate on some of the big issues and trials in cardiovascular (CV) medicine that seem to be coming up in 2019 and beyond.

Again, my guest today is my friend and colleague Mike Gibson, who is a professor of medicine at Harvard Medical School in Boston. Mike is also an interventional cardiologist at the Beth Israel Deaconess in Boston, and finally, he is the CEO of the not-for-profit academic research organization Baim Institute for Clinical Research. Mike, thank you for joining us on

C. Michael Gibson, MD: Thank you for having me back, Bob.

Harrington: We had a lot of fun reflecting back on 2018, but now I want to look forward to 2019. Here are some of the things I want to run through: wearable technologies to pick up atrial fibrillation (AF) and other physiologic measures, aspirin for secondary prevention, artificial intelligence (AI) and machine learning, open-access and open-source publishing in the European Union, drug pricing, and CV disease as a cause of death and disability in the United States. That is a lot to talk about, Mike. Let's jump right in.


Harrington: What are you seeing on the horizon in the wearable space?

Gibson: For 17 years, I worked on implantables, and I will be presenting the results of a trial using an implantable device to detect ST elevation at the CRT meeting coming up. Wearables are so much simpler: You strap the device onto your wrist, and you get heartbeat and rhythm.

It's important to note that the Apple Watch, for instance, is a consumer product. It is not labeled as a diagnostic medical device. But it is going to be used at scale—I know you are looking at over 400,000 largely young people [in the Apple Heart Study].[1] We are looking forward to hearing what you learn.

A lot of us have tremendous interest in how to detect "AF" in older people. What is AF? How long do you have to have AF for it to have an impact on outcomes? We have so much to learn from the Apple Watch and technology like that. I am looking forward to seeing what you guys have to say in 2019.

Harrington: The Apple Heart Study is a project led by Ken Mahaffey, Marco Perez, and Mintu Turakhia, working with a group from around the country to try to understand the role of a wearable device in detecting AF. These wearable devices can detect a lot of physiologic signals that can then be transformed into the probability of it being something like AF. It raises a whole set of questions. What do you do when you detect AF, or the probability of AF, in a background population that is asymptomatic? Do we redefine what AF and its risk might be? How do we manage people who have AF with these wearable devices?

There are a lot of questions and a lot of exciting work to do. I, too, look forward to seeing Ken, Marco, and Mintu present the first observations from their study sometime in early 2019.

Aspirin for Secondary Prevention

Harrington: It's a bit of a detour, but let's talk about the ADAPTABLE trial.[2] I have been involved in this trial with our Duke and Vanderbilt friends and a variety of other places around the country. We are looking at two different doses of aspirin for secondary prevention. It is a large pragmatic trial funded by Patient-Centered Outcomes Research Institute (PCORI), and all done through electronic health records (EHRs) in about 40 centers throughout the United States.

Aspirin for primary prevention was not a winner in 2018. What are your views on aspirin for secondary prevention, and what do you think of aspirin 81 mg versus 325 mg and the outcome? Also, what do you think about doing trials through the EHR?

Gibson: For secondary prevention, a lot of data show that there is a complicated use relationship, with 75 mg seeming to be the sweet spot. Higher doses give you higher bleeding rates and not as good efficacy as 75 mg, and then if you go below that, you have higher thrombotic rates.[3,4] I will be interested to see what you find. I imagine the low dose will be better.

I do think the future is exactly what you are doing by engaging patients directly, cutting out a lot of the cumbersome data collection that occurs with paper-based case report forms, and making it all live and selectable through mobile devices. That is going to be the future. I congratulate you on all of the efforts.

Harrington: The assumption for secondary prevention with aspirin is that 75 mg has similar efficacy for most people—patients with diabetes may be a notable exception—and you get less bleeding with the lower dose. We will finish enrollment in 2019, and results will follow from there.

I am with you that the whole concept of engaging our patients in the research process as real citizen scientists is critical. Using new technologies will allow us to avoid having to collect data in a duplicate way that we have been doing for years. To me, this has got to be a winner.

Artificial Intelligence

Harrington: Speaking of new technology, I know your lab is spending a lot of time thinking about the tools of AI, including machine learning and neural networks. How do you see this playing out? You cannot pick up a journal these days without seeing a machine learning paper in it. Some of it is hype, some of it raises interesting methodologic questions, and some of it is offering up new opportunities. For example, a machine helped me read a chest radiograph.

Gibson: I think the hype is probably, and sadly, some of the expert decision-making capabilities. We are just not seeing that big of an advance using machine learning over very well-done logistic regression in decision-making. Logistic regression tells you about the population in general, whereas machine learning talks about this one patient in front of me. And it seems to be better calibrated if the technology is deployed over and over, but not perhaps quite as good as we are.

In imaging, we are going to see tremendous advances. In sifting through genetic data and pairing that with wearable data, we are already seeing tremendous advances. There is the United Kingdom Biobank study, where they put pedometers on 91,000 people and found 14 new genetic loci that determine exercise and sleep.[5] That is a great example of combining wearables with AI and machine learning. I do think there will be a lot of this in the future.

ZIP code and socioeconomics are sadly playing a much bigger role in health than they should.

Harrington: There is no question that with cloud computing and our ability to obtain massive amounts of very disparate data (eg, continuous activity data, imaging data, genetic data, demographic data), conventional statistical methods have a hard time with very disparate types of data. Perhaps some of these advanced methods of computation will allow us to gain insight into the areas we did not have before. I am excited about it, but we do have to be aware of the hype.

Gibson: In mining all that data about geospatial location and particulate matter, your health is dictated so much by your ZIP code now. I think it allows us to integrate all that information.

Cardiovascular Disease and Socioeconomics

Harrington: I thought one the most important papers in late 2018 came out of a group of investigators here at Stanford (not anything I was involved with). It gets to the exact topic you brought up. What are the leading causes of death in the United States? For many decades, it has been cancer and heart disease, with heart disease, unfortunately for those of us in the field, being number 1. By 2023, it looks as though heart disease is going to be passed by cancer as the leading cause of death in the United States.[6] Interestingly, it is going to be very geographically dependent.

Your comment about ZIP code is critically important here. If you are from a lower-socioeconomic, lower-income area, it looks as though heart disease will still predominate because of things like smoking, environmental exposure, lack of exercise, obesity, poor treatment of hypertension, et cetera. If you are in areas with higher socioeconomics, perhaps the leading cause of death will be cancer.

Gibson: We are looking at trade-offs in the way you die, based on your socioeconomic status. If you did a competing risk analysis, I think what we are doing is preventing those deaths from CV disease in the higher socioeconomic classes that only allows them to live so that they can die of cancer.

Some people think we won the war against heart disease. We have not won the war against heart disease at all. We have had a few battles here and there. It is sad to see the investment in heart disease research going from about 20% of research and development dollars down to about 5%-10% now when it continues to be such a big killer.

It is not only what you eat; it is actually where you eat as well. ZIP code and socioeconomics are sadly playing a much bigger role in health than they should.

Harrington: From a global perspective, heart disease remains the global leading cause of death and disability, and we clearly have not won this war. We need to do better. We need to concentrate on cigarette smoking, poor blood pressure control, diabetes, obesity, hypertension, and lack of physical activity.

Open-Source Publishing

Harrington: The next topic is open-source publishing and what is going on in the European Union.

Gibson: Something that is really coming on strong now is a movement that is called Plan S, also known as cOAlition S.[7] This is a global movement, which started in Europe, that would require the distribution of free, open-source medical information if you have received money from any kind of government to fund the research. It would switch the traditional academic subscription model to some other model. You would not have subscriptions; you would not have reprints and permission as revenue.

If you have a hybrid model where some articles are open-access and others are behind a paywall, that will not meet the requirements of Plan S. Those hybrid models would maybe have a cap. The number being floated out there is that it would be $7000 to publish your article. I think there are going to be new models of how we fund publishing of research. This is a very big development in the publishing world.

Harrington: You and I have certainly opined on the whole issue of data-sharing about how quickly you need to share your data following a completion of a set of studies, and whom should you share it with and the methods that that involves. This is all a part of that story. It is an exciting story, because I think the intent is to get data out there, and shared and worked on by a lot of individuals with disparate skill sets. Clearly, there are a lot of issues for us to think about.

Drug Pricing

Harrington: The 2018 elections in the United States showed that healthcare was a dominant topic. There is no question that both political parties have gone after this notion that US drug prices are too high.

Gibson: They are very high, and this is something that both parties can clearly agree on. I think the solution, though, remains unclear. It is frustrating to see that in the United States, the cure for hepatitis C can cost $50,000-$80,000, but that same drug in India is priced at $1000.

One potential strategy is to bring greater parity around the world in drug pricing. A lot of concern has been expressed about the United States paying a lot for the development and then paying a lot for the drugs that are developed here. Maybe those costs could be spread around a little more throughout the world.

From my vantage point, I cannot tell what they [pharmacy benefits managers] do other than make money for themselves.

Harrington: It certainly is a challenging topic. The policymakers and others in society have to deal with this whole notion of the middleman, the pharmacy benefits managers. At least from my vantage point, I cannot tell what they do other than make money for themselves. How are we as a society going to continue to be able to allow this, while consumers are paying way too much for drugs and people are charging way too much for drugs? There does seem to be this group in the middle that is benefiting unfairly.

Gibson: Yes, the pharmacy [benefits] managers will have to go. On the other hand, think about the days when statins were pricey. Now you can take a statin for $3 a month, beta-blockers are pennies, aspirin is pennies. I think we can get to very affordable drugs for the masses of people once we get through the period of exclusivity.

Harrington: The proprotein convertase subtilisin-kexin type 9 (PCSK9) inhibitors in 2018 were a good example. Here, we actually had formal cost-effectiveness analysis[8] based on clinical trial results that basically identified what a reasonable charge would be for these drugs. Because there has been such slow uptake of the drug, the company finally had to react to that and lower the price of their drugs down into the zone that would be considered cost-effective.

Gibson: Yes, it is a moving target. The numbers have come down to be much more compatible than, say, the $3000 price for Crestor. Now we are looking at a $5000 price tag, so much more cost-effective. Some of the future strategies will probably be to make it less costly to manufacture these agents and allow a lower price point.

Harrington: Mike, this has been fun. First, we reflected back on 2018, and now we are looking ahead. It is always fun to catch up with you and to see what you are thinking about.

My guest has been Mike Gibson, professor of medicine at Harvard and CEO of the not-for-profit Baim Institute in Boston. Mike, thank you for joining me on Medscape Cardiology.

Gibson: Thank you for having me back.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: