Latest / Coffee People Podcast / Dr. Gregory Marcus, Associate Chief of Cardiology for Research at UCSF Health

Transcript

Dr. Gregory Marcus, Associate Chief of Cardiology for Research at UCSF Health · Coffee People Podcast

0:24
None. We're officially live on the
0:29
Coffee People podcast that I'm glad you're here, Doctor, if you
0:34
wouldn't mind for anyone listening today, could you
0:36
introduce yourself Dame and and what it is that you do?
0:41
Yeah, so I'm so thanks so much for having me.
0:44
Delighted to to be here. So I'm Greg Marcus.
0:47
I'm a what's called a cardiac electrophysiologist, which means
0:52
I'm a sub specialist within cardiology with a focus on heart
0:57
rhythm disorders. So I do pacemakers and
1:01
defibrillators and and catheter ablations.
1:04
But these days, although I have that practice at the University
1:09
of California, San Francisco, I spend most of my time doing
1:13
clinical research, a lot related to clinical trials and
1:18
epidemiology, which has led me to quite a bit of research on
1:24
common exposures such as alcohol and obviously pertinent to this
1:29
particular discussion, caffeine and coffee.
1:33
Sure, I just have a cardiology question before we get too far
1:38
down the road. Not that I'm asking for medical
1:40
advice, but a few years ago I had to wear this device to track
1:43
my heart rate. Do you still have to shave your
1:45
chest or is that all automated and Bluetooth now?
1:48
You still need to shave your chest unless you you know if you
1:52
have an Apple Watch you can report intermittently.
1:55
Good to know. But yeah, so we we are here
2:00
'cause it's a coffee show. I stumbled on this paper doing
2:04
some light reading called Acute Effects of Coffee Consumption on
2:07
Health Among Ambulatory Adults, and that's gonna be part of what
2:11
we're discussing today. But I'm wondering before we get
2:14
there, how did you end up getting started down this path?
2:18
You know, did was coffee part of your life before it became into
2:21
your life as as a medical study? Do you have any first memories
2:26
of coffee you'd like to share? Yeah, so actually I think it's
2:30
partly because I can't. I, I am a, a big coffee fan and
2:35
a regular coffee drinker, but perhaps ironically, I, it may be
2:40
because I came to it late in life that it became a research
2:45
interest because so I went through medical school and
2:49
residency and fellowship. And this is before the modern
2:52
day of, of work hour restriction.
2:55
So, so these days, for example, a resident is not supposed to
2:58
work more than 80 hours a week. We didn't have that in the, in
3:02
the, as we say, the days of giants when I was training where
3:06
we would be on call, you know, overnight the next day,
3:10
sometimes every third night without a break, sometimes every
3:14
other night and, and all weekend.
3:17
I, I didn't have any coffee throughout that.
3:19
And then we had babies and I was, my wife and I were really
3:23
5050. I, I was the primary diaper
3:26
changer. And even when she was nursing, I
3:29
would be the one to get up to pick up the baby from the crib
3:32
and, and bring one of our daughters to her.
3:35
So was awake a lot, sleep deprived a lot, no coffee and it
3:39
wasn't until and I can't say why, although for some reason my
3:44
wife was was similar and then she started to drink these
3:51
concoctions with this foam that looked just delicious.
3:56
And this is in my probably early, early 40s.
4:00
And I thought, man, that just that just looks really tasty and
4:03
that so my introduction was via the foamy drinks and then it was
4:08
cappuccinos. So I'm AI remain a huge
4:10
cappuccino fan. And then when the pandemic hit,
4:15
I was like, oh, man, I can't, you know, can't get my
4:18
cappuccino. I guess I have to learn how to
4:20
just make some coffee. And and now I'm, you know, a, a
4:24
twice a day drinker. Love the taste.
4:28
Favorite coffee shop is actually it's called Cafe Central.
4:32
It is in the University of California San Francisco library
4:38
and if anyone's ever, you know, walking by there, it's it is
4:41
publicly because it's a public institution.
4:43
It's at least ostensibly publicly available.
4:47
So I would recommend recommend that place.
4:49
I like that medical school didn't break you, but having a
4:53
baby dead? Well, it was after, it was
4:56
actually even after the babies. It was purely for the taste, but
5:01
yeah. So she had started.
5:04
OK, so I, I mentioned that study that we're talking about.
5:08
One thing I realized that I didn't quite understand as I was
5:12
kind of scrolling through it and, and trying to make sense of
5:14
it all I should warn you that I failed science in in high school
5:19
is that I didn't really understand what a clinical trial
5:22
was. I'm wondering if you could give
5:24
us just some basics. How does a you know, how does
5:27
someone have an idea that says I want to study this thing or test
5:30
this idea and then run it through some sort of program to
5:36
a conclusion? Yeah.
5:38
Well, I think it may be helpful to draw the distinction between
5:44
what we tend to call observational studies versus
5:48
more interventional clinical trials.
5:51
So the great majority of research that's published, and
5:55
this includes the great majority of medical research and
5:58
definitely the majority of research on coffee and other
6:01
common substances falls into this category of observational
6:07
meaning. You, for example, you ask a
6:10
bunch of people how much coffee do you drink?
6:13
And then you, so then you have a record of that, of some number
6:17
of people and then you follow them over time and you see what
6:21
happens And you see, oh, these people developed a heart attack.
6:24
These people developed a heart, an abnormal heart rhythm.
6:28
And that can be very useful. And it tends to be quite
6:32
feasible and easy to do, which is why they're so commonly done.
6:39
But they are prone to what we call confounding, meaning
6:45
there's a reason that some people are drinking more or less
6:48
coffee. There's there are different
6:50
characteristics there. And it could be that that other
6:53
thing that is determining why they're drinking more or less
6:57
coffee is the true causal actor. That is what's actually
7:01
responsible for differences in heart attack or stroke or
7:05
whatever your cancer, whatever your outcome is, sure,
7:08
depression, for example. So you can try to address that
7:12
by doing statistical analysis that take those confounders and
7:16
you adjust for them in a statistical model.
7:21
But that is imperfect for a bunch of reasons.
7:24
One is we never really fully capture those things perfectly
7:29
well. And this is not going to be an
7:32
exhaustive list, but another really important issue is you
7:36
have to know what those confounders are.
7:39
And often times we have no idea, right?
7:41
We just don't know what they are.
7:42
Sure. So really the only way to
7:46
confidently infer a causal relationship is ideally to
7:53
randomly assign your whatever your exposure of interest is.
7:58
In this case coffee. If it's randomly assigned, then
8:03
you don't even need to know what the confounders are are they
8:06
kind of shake out in the randomization.
8:09
You will end up with about half the people randomized to one arm
8:14
versus the other with, you know the same kind of
8:17
characteristics. So that is more of an
8:20
interventional clinical trial where it is looking, you know
8:25
you are, you are the one introducing your exposure
8:30
coffee. We've done this in a couple of
8:33
ways. So one is and and I can talk
8:36
about we kind of have two major trials on this.
8:38
We've also done observational research that we can talk about.
8:43
And in one trial we really did randomly assign half the people
8:49
to go ahead and consume caffeinated coffee at least a
8:52
cup a day versus try to avoid all caffeine.
8:56
This was our what's called our decaf trial.
8:58
Does eliminating coffee avoid fibrillation?
9:02
Fibrillation in this case being this very common arrhythmia.
9:06
I do love an acronym. And then the other one which was
9:11
which is Crave which stood for coffee and real time atrial and
9:16
ventricular ectopy which are early beats that we all have and
9:21
we can talk about that atrial ectopy coming from the upper
9:25
Chamber of the heart, ventricular ectopy coming from
9:27
the lower chamber. And in that study, we actually
9:31
people were randomly assigned day-to-day to avoid versus
9:35
consume caffeinated coffee, so. They're not switching between
9:39
the two. Switching between the two, but
9:41
again, very importantly in a random order, because if you
9:45
just followed people who you know, they consumed more coffee
9:50
on Monday and less on Thursday. Well, what what's going on on
9:53
Monday and Thursday that could really be responsible for that.
9:58
And there there's an advantage in that each person serves as
10:02
kind of their own control, which gives you more power.
10:07
But in that case, you're limited to really looking at very short
10:10
term effects. Sliding into your screen is not
10:16
just any coffee can, it's a Rostar coffee can.
10:19
You don't need thousands personalize, just one in
10:23
Rostar's design lab. No design experience needed.
10:27
Choose a pattern, add your logo and text, or create something
10:31
memorable. With AI being made in the
10:34
Midwest, you can expect it on your doorstep in no time.
10:38
Now that was smooth. I guess one thing I, I don't
10:45
quite follow just from not being in that world at all is
10:50
somebody, I'm assuming you were involved in this, but someone
10:53
has this idea. How do you turn that into a real
10:56
thing or how do you get that funded?
10:58
Who do, who do you go to and say, hey, can I have permission
11:01
to do this? Do you need permission?
11:02
You know, those are all questions that are floating
11:04
around in my head. And on the other side of that,
11:07
deciding upfront that there's enough value to do it.
11:11
Yeah, great questions. So I, I, I think there are three
11:16
questions in there. One is kind of what motivated
11:19
pursuing this research. 2 is what kind of permissions do you
11:22
need? 3 is where do you get the
11:25
resources and the funding to support such a study?
11:29
Great, very important questions. And, and I would say, you know,
11:33
all of this stuff, by the way, applies very broadly to the
11:36
scientific process, which I think is frankly super important
11:42
to the flourishing of us, of, of us humans.
11:45
So I think this is a really important topic.
11:48
So specific to coffee, you know, what kind of what motivated the,
11:51
the question. So there has long been, as I'm,
11:55
I suspect your listeners all are, are well aware this
12:00
conventional wisdom that caffeine and caffeinated coffee
12:06
cause abnormal heart rhythms. They can make your heartbeat
12:11
abnormally beat too fast, etcetera.
12:15
And so, and this is not just a layperson sort of understanding,
12:20
this was operative in clinical medicine.
12:23
So if someone had a diagnosis of a disease, specifically in this
12:29
case atrial fibrillation, that's very, very common, in fact the
12:33
most common sustained abnormal heart rhythm that affects more
12:39
than 10 million Americans like right now.
12:43
And it's a really important 'cause of stroke and can can
12:46
make you feel really crummy, has a lot incomplete to heart
12:49
failure, a lot of other problems.
12:51
Was essentially the norm for practicing physicians to tell
12:57
their patients with atrial fibrillation.
12:59
Oh yeah, yeah, you should avoid caffeine and coffee based on
13:03
essentially 0 evidence. Just again received wisdom.
13:08
Oh yeah. We just kind of already know
13:10
this, but no one had tested it rigorously.
13:14
And that sort of phenomenon where there's conventional
13:17
wisdom without evidence, that's a great area to research.
13:23
And often, sometimes we confirm the conventional wisdom, but
13:26
sometimes we upend it, as I think we have done with atrial
13:30
fibrillation. So in terms of getting
13:34
permission and and getting resources.
13:36
So a lot of this has to do with the experience of the
13:41
investigator. So just as we go through medical
13:43
school and residency and fellowship there, there are
13:46
there's formal coursework one can take in doing research and
13:49
of course a lot of experience to be gained and knowledge to be
13:53
gained over time as one participates initially, usually
13:58
as a mentee or as a trainee under the supervision of a
14:03
experienced investigator. And then you kind of work your
14:06
way up from someone working on someone else's ideas to
14:11
generating your own ideas and having your own team, usually
14:16
with biostatisticians. And again, there's a knowledge
14:20
base here that is related to, but somewhat separate from
14:23
medicine that has to do with statistics, epidemiology, the
14:28
study design, sorts of things that I have been fortunate to,
14:33
to gather over time. I also did as a junior faculty
14:36
member, some coursework. I got a master's in, in advanced
14:40
studies and clinical research at UCSF, which helped a lot and,
14:45
and got involved in, in conducting research.
14:48
And as with anything, the best way to, to learn it is to just
14:52
start to do it and fail and learn from your, you know,
14:55
mistakes and failures. And so that's been building over
14:59
a long period of time. And for sure if we are going to
15:03
enroll anyone, any person in any sort of study, we need
15:09
permission. And the nature of the permission
15:14
depends on the study. But generally speaking at the at
15:19
the least, we need what we call Institutional Review Board or
15:23
IRB approval. So all major academic medical
15:29
centers or, or institutions that conduct research, they will have
15:33
an institutional review review board with a set of processes
15:38
and people who serve on committees that will review
15:44
these applications that have a certain format.
15:46
And there's often times it's, it's almost unheard of that you
15:50
would, for example, submit an IRB and then yeah, they just
15:54
approve it. Almost always there's feedback,
15:56
there are questions, there's iterations and we revise and,
16:01
and respond and ultimately certainly cannot enroll.
16:06
We are not allowed to enroll a participant in any study until
16:10
we have that higher B approval. And part of that process
16:13
includes drafting, crafting, refining a consent form for
16:21
every research participant. So this is again, when we're
16:24
doing a clinical trial, if we're going to be conducting something
16:28
prospectively, meaning we're not, you know, we're going
16:30
forward in time with that person who is enrolling, they need to
16:35
provide informed consent for, for the observational studies I
16:42
described, you don't necessarily need that because it's often
16:45
just not feasible. But that's an example of why
16:49
those studies are easier to do and less expensive, right?
16:52
If if you go backwards in time and you have some data set or
16:56
you're just looking at medical records, you don't have to spend
16:59
the time enrolling those people, interviewing them.
17:02
You're not delivering an intervention.
17:03
You're just looking back to see what happened to them.
17:06
Sure, and just in that in that comparison though, the
17:11
observational studies don't control as many variables and so
17:14
their results are maybe not as precise.
17:17
That's right, that's right. They're they're methodologically
17:21
not as good and, and technically and, and I think this is really
17:24
important, especially as we all consume media, which is often
17:30
the the communicators of the results of research.
17:34
Technically speaking, that sort of observational research is
17:38
really hypothesis generating. So you you, you see some, you
17:43
know, headline coffee is associated with improved
17:48
cognition. This is a recent study, a good a
17:51
good study methodologically careful, but it was not a
17:54
prospective randomized trial, technically speaking, to be
17:58
rigorous about it. That is hypothesis generating it
18:01
it presents the hypothesis and provides some evidence in favor
18:04
of it that drinking caffeinated coffee improves or or prevents
18:10
the decline of cognition. But it's really the when we
18:13
again, when we want to say X causes Y, that is really gleaned
18:18
from randomized trials. So like randomized trials.
18:21
So, you know, when we talk about, we have strong evidence
18:24
that statins prevent heart attack, those are from
18:28
randomized, really double-blind, placebo-controlled trials.
18:32
People with generally with risk factors for heart attack
18:35
randomly assigned, they experience less heart attack if
18:39
they are given, given the statin versus the placebo, for example.
18:44
And I, I should mention for those observational studies, you
18:48
also do need IRB approval for those.
18:50
So even if you're not obtaining consent, you need to have this
18:53
objective third party review. Again, you need to respond.
18:58
It all needs to be done according to certain ethical
19:01
standards, you know, doing everything we can to to make
19:05
sure that there's no issues with privacy, that everything is as
19:09
protected as possible, that sort of thing.
19:12
And then to get to your third question about funding.
19:16
So this is a ongoing challenge or struggle for investigators.
19:24
And so there are a number of potential sources of funding.
19:29
The most prestigious frankly and arguably the best is, is
19:36
taxpayer funded research such as from the National Institutes of
19:41
Health that is extraordinarily difficult to get.
19:46
It's a very rigorous process. And I think that all it should
19:50
be, I mean it, it, it should be a very high bar and it that
19:54
requires really critical peer review.
20:00
Again, as it should, the advantage there is there's not
20:04
really, there's not a vested interest in the outcome.
20:09
So another source is industry funding, which can be useful.
20:14
You know, some company wants to support research to do something
20:19
there. There's a concern about
20:22
conflicts of interest, right? The company understandably wants
20:26
to promote whatever their product is.
20:28
And so one of the there, there are many benefits of taxpayer
20:33
funded research and there's other forms apart from the NIH
20:36
different some states have it. So in California we have a
20:40
tobacco research program that's funded by taxes on cigarettes,
20:45
for example. That was something the public
20:47
voted for. But the advantages are again,
20:50
you get peer reviews. So they, the NIH or those
20:53
tobacco group, they will invite the experts in the field who are
21:01
generally in academia to come serve as peer reviewers.
21:05
So it is again, rigorously peer reviewed and again, almost
21:10
unheard of to submit a grant and even the best scored grants just
21:14
to say, yeah, you're funded. Almost always there's a lot of
21:17
critique, multiple reviewers. You need to respond to them.
21:21
You need to revise your your plan and you need to then
21:25
satisfy them. So again, it's a very difficult
21:28
process, but that's good. I think this is a good
21:31
application of meritocracy, which works well for it helps to
21:36
assure, you know, really rigorous science is being
21:39
conducted. And then I mentioned industry
21:43
and then a third. And these are not, again, not
21:45
necessarily exhaustive, but in in big buckets.
21:48
Another source is philanthropy. So which all of these things
21:54
have their pros and cons, but some folks with a lot of means
21:59
sometimes want to give back. And so they will donate funds to
22:05
either foundation. So like the Gates Foundation
22:07
does a lot of work on malaria and, and things like that.
22:12
And then there are some private donors that we'll give to, for
22:16
example, the division of cardiology.
22:17
So these studies were well. So crave as I just described,
22:23
the randomized case crossover trial was actually funded.
22:27
It's, it's a, it's a bit of a hybrid where philanthropy led to
22:34
some funds that our division chief transformed into a
22:39
competitive essential sort of proposal.
22:46
It's called the Cardiology Innovation Award.
22:48
So it's competitive, you need to it's peer reviewed within UCSF,
22:54
but the the source of the funds arose from philanthropy.
22:59
It sounds like they're. So maybe I'm misunderstood.
23:03
Not maybe I'm just because I'm not in this world at all.
23:07
How much of your job is actually working on the trials or
23:10
exploring these health things, and how much of it is trying to
23:13
get the money? Yeah, it's a great question and
23:16
it varies from time to time. Unfortunately, too much time is
23:21
spent writing grants because it's so competitive that most of
23:25
them are rejected. And it varies.
23:29
I mean, it can, but I would say on average maybe 2030% of my
23:33
time as a fairly senior investigator now spent writing
23:38
grants trying to get funding. Sure how, yeah.
23:41
And I'd say even for me, I'm in my 40s until COVID, really I
23:46
didn't wasn't aware of the NIH or what it did.
23:49
And now I think people are much more aware because it's in the
23:52
news a lot, especially the last couple of years, not always for
23:56
great reasons as far as funding cuts and that sort of things.
23:59
And and we're trying to certainly figure out what the
24:02
value is there as a society. I think some of us feel one way
24:05
and some of us feel a different way.
24:09
I want to get back to that, to the coffee part of this, which
24:12
is in at least I have a little note here.
24:15
And you know, the conclusion, A quoted sentence that I wrote
24:18
down from that first paper, the effects on the effects of coffee
24:23
consumption, it says that in, in this randomized trial, the
24:26
consumption of caffeinated coffee did not result in
24:29
significantly more daily premature atrial contractions
24:34
than the avoidance of caffeine. And that's a great sentence.
24:39
I, I feel like I, I felt like I knew it.
24:41
And then the more I thought about it, I started to realize,
24:43
well, I don't know what this means.
24:45
Does this mean that I that I can just drink whatever I want?
24:49
I can drink as you know, I can drink a whole pot of this coffee
24:52
every day or every hour or, you know, is.
24:54
So when you write that down, and I think intentionally, it has to
24:59
be very narrow to fit the scope of what you're doing and also be
25:06
very specific. But what is what is a conclusion
25:09
in in that trial or in the decaf trial?
25:11
What did, what do those mean to coffee people, people who are
25:14
out here drinking coffee, just normal people who don't
25:16
necessarily read these papers or understand, you know, the whole
25:20
process behind them. Yeah.
25:22
Happy to answer that. So it might help to maybe
25:27
describe the trial and what we found.
25:29
I can hone in on that particular sentence where the journal and
25:35
the editors of the journal also have a bit of a role to play as
25:38
I will describe. So in Crave, we essentially
25:43
enrolled about 100 healthy volunteers who consumed
25:47
caffeinated coffee on a regular basis and we're willing to go
25:51
without caffeinated coffee, which was which made it a little
25:56
hard to enroll. There were many people like no
25:58
way I can go without my caffeinated coffee for a day or
26:02
two. And we gave them all
26:04
continuously recording EKG patches similar to the one it
26:08
sounds like you wore. We gave them all Fitbits to
26:12
record their step counts and their sleep.
26:15
We gave them all continuously recording glucose monitors and
26:20
then and their study participation was essentially 2
26:24
weeks. That's about the time that that
26:26
EKG patch will stay on and they were randomly assigned every
26:32
night 8:00 PM. They were told, OK, tomorrow,
26:37
either they were told drink all the caffeinated coffee you want
26:42
or they were told try to avoid all coffee and caffeine.
26:48
And then the next morning 8:00 AM, they got a reminder.
26:52
Remember today you can, you know, do X or avoid all
26:57
caffeinated coffee and caffeine. We asked them to press so that
27:03
EKG recorder has a button on it that will time stamp an event.
27:08
It's used clinically to we tell patients, hey, when you have
27:11
your symptoms of your heart rhythm disorder, press the
27:15
button so we can line up their symptoms with what's going on
27:17
with their heart rhythm. For this study, we asked them to
27:21
press the button whenever they had a cup of coffee so we can
27:25
time stamp when that happened. We also, so they downloaded this
27:31
app and I should mention I helped to run a digital research
27:35
platform called Eureka that was originally funded by NIH and it
27:39
was it's designed to facilitate mobile health research, meaning
27:44
research that uses mobile apps and sensors and wearables.
27:50
And one of the things, and again, this is of course, IRB
27:54
approved and everyone provided informed consent, including the
28:01
permission to follow their geolocation for visits to coffee
28:06
shops. So the way we validated whether
28:10
they really did follow their randomization instructions was
28:13
several fold. So #1 as I said, they pushed the
28:16
button in real time. So we knew whether they consumed
28:21
or not. Every morning that eight with
28:23
that 8:00 AM reminder, they were asked to fill out a survey.
28:27
And we asked them in that survey that was delivered digitally.
28:31
Did you actually have any coffee yesterday?
28:34
And if so, you know how much? And there was really no
28:37
incentive for them to lie. It's not like they were going to
28:39
get paid if they did or didn't, Although we understand, you
28:42
know, there's some social stigma potentially if if they feel like
28:46
they're not kind of following instructions.
28:48
So then we also Geo fenced visits to coffee shops and then
28:53
we also told the participants we will pay for all of your
28:56
caffeinated coffee as long as you give us a date stamped
29:01
receipt and whether you were supposed to or not.
29:04
So we kind of motivated them to give us some objective evidence
29:08
and it turned out that the compliance with their
29:11
randomization assignments were not perfect.
29:14
We're all human, we understand that.
29:16
But they were really good. So most of the time all the
29:19
evidence suggested yes, they were really following the
29:22
randomization instructions. I feel like it's dangerous to,
29:25
if you were to collect your patient or your, your trial
29:29
participants and tell them that you would buy them coffee.
29:33
I feel like on the days that I could drink coffee, I'd be like,
29:36
all right, it's 7:00 AM, let's go.
29:38
Yeah, yeah. So, and we also actually, we
29:44
haven't analyzed this yet, We did ask everyone at 8:00 PM, how
29:48
do you feel right now? Like 'cause we wanted to
29:49
understand about their mood. And what's interesting is we
29:53
realized later they got that question after they were told
29:57
what their instructions were for the next day.
29:59
So this is a plan to, so we can see like how bummed people were
30:02
to learn, Oh no, I'm not supposed to have coffee tomorrow
30:05
versus how happy they were to learn that they should.
30:09
So the primary outcome for this trial was this, the premature
30:14
atrial contraction count, which you alluded to in reading that
30:17
sentence. So premature atrial contractions
30:21
are early beats that come from the upper chambers of the heart.
30:25
The upper chambers of the heart are called the Atria.
30:28
So we also call them PA CS. We all have PA CS.
30:33
You put a monitor on any, essentially any adult walking
30:36
down the street, you will find PA CS.
30:39
Some people have more than others.
30:42
We and others have previously shown that people with more PA
30:46
CS have a higher risk of atrial fibrillation.
30:51
And in fact, when we do procedures to get rid of atrial
30:55
fibrillation, one of the things we're doing is probably
30:58
eliminating PA CS that can trigger this abnormal rhythm.
31:03
So the idea was we wanted to enroll healthy volunteers, but
31:08
how could we capture a something that is clinically relevant in
31:14
the general population? And so we landed on PA CS as
31:18
something again, we all have, and it's a what we call
31:21
continuous outcome, meaning it's, it's not a yes or no,
31:24
it's, it's a number which helps with statistical power, gives
31:28
you a lot of statistical power, especially because you can
31:32
compare numbers of PA CS on versus off again, within the
31:35
same people. Now we did look at other
31:40
outcomes, which I will describe in a second, but there there is
31:47
a standard in in clinical trials where generally we want to pre
31:54
specify our primary outcome. And the journal where this was
31:59
published, the New England Journal of Medicine is
32:02
appropriately very persnickety about making sure you really
32:08
hone in on that primary outcome, especially when it comes to the
32:11
conclusions, which is totally fair, right?
32:14
Because you could have 10 other outcomes and just by chance
32:18
alone, you might find one that meets your statistical
32:24
significance threshold. But really that's just from
32:27
chance. So if you could just like Willy
32:29
nilly pick which, oh, this one was positive, but you know, then
32:34
that's, that's evaluating things after the fact, which enhances
32:39
the possibility of a false positive.
32:42
So, so hence the concluding sentence was focused on that
32:48
primary outcome, which for this particular trial was just PAC
32:52
counts. And yes, indeed, we found no
32:55
relationship between caffeinated coffee consumption and PAC
32:59
counts. So that's not a perfect
33:01
surrogate for atrial fibrillation, but it's pretty
33:06
good. And the prior observational
33:09
evidence again that of inferior methodology also suggested
33:14
actually that caffeinated coffee did not seem to increase the
33:17
risk for atrial fibrillation and may even protect against atrial
33:23
fibrillation. And we can talk about that a
33:24
little bit more when we get to to decaf.
33:27
Sure. And I just kind of want to make
33:29
a note here for anyone listening.
33:30
I think what I'm what I'm hearing is just someone hearing
33:34
from you your study. It builds on other studies like
33:41
a clinical trials building on an observational study, and you're
33:43
essentially looking to create whether it's just you or a group
33:47
of researchers or over time. A indicator that we can say this
33:53
feels reliable as far as information we put into the
33:57
system, into the medical complex, is that.
34:00
That's right. OK, So we apply, you know,
34:03
there's this the Bayes theorem, which has to do with, you know,
34:08
what is the likelihood based on the prior evidence of a
34:13
particular outcome. And so we had this hypothesis
34:16
that itself, it wasn't just pulled out of the air, it was
34:20
based on a lot of other prior research, albeit all
34:24
observational, that caffeine and especially caffeinated coffee
34:30
did not seem to increase the risk for this very important
34:34
heart rhythm disturbance and may even protect against it.
34:38
We thought about, well, how could we test that in a
34:39
randomized way and came up with this CRAVE trial.
34:44
Now some other outcomes, again, they were not the primary, but I
34:49
think they're still interesting and they themselves will
34:52
hopefully generate more research and and replication is also very
34:57
important before we make any firm conclusions.
35:01
So one was that there were actually more PVCS.
35:05
So these early beats coming from the lower chamber premature
35:09
ventricular contractions, which we all also have, were actually
35:15
increased on days when people were randomly assigned to
35:18
caffeinated coffee. And also the more caffeinated
35:21
coffee they consumed based on those button presses, for
35:24
example, the more of those early beats from the lower chamber
35:28
they had. So PVCS may, if they are very
35:34
frequent, can increase the risk of weakening of the lower
35:40
Chamber of the heart. So we actually have So a, we've
35:44
talked before about funding. I have a grant that I've
35:48
submitted that took me many, many months to put together that
35:52
is going to be reviewed in about 10 days by the NIH to try to
35:59
understand this relationship between caffeinated coffee and
36:02
premature ventricular contractions to #1 duplicate
36:06
those findings. But in this this case as the
36:09
primary outcome and also to understand is this really
36:12
matter, Is this relevant to clinical medicine, Is it
36:16
relevant to people with a lot of PVC's where this is is most
36:20
important? I would mention that
36:22
observational data if anything seemed to suggest that
36:27
caffeinated coffee consumers have a lower risk of heart
36:30
failure. So that's good.
36:32
There's some suggestion that maybe at very high doses of
36:36
caffeinated coffee that risk may go up a little bit, but most
36:39
people who drink caffeinated coffee seem to have a lower risk
36:42
of weakening of the heart. How often when you finish a
36:47
study and you're putting this information out there into the
36:49
world, is it, how often do you see the results or the
36:55
conclusions get skewed as they are filtered through the
36:59
perspective of a, a media person or a somebody from an industry
37:04
that you're studying that may be affected?
37:06
Or is that something that you have to let go of or that you
37:11
continually try to battle and say, no, this is actually what
37:14
that means. Don't jump to this further
37:16
conclusion just because of what I've put out in the world.
37:19
Yeah. So it depends a lot on the
37:24
audience and on the findings. So if the audience is a medical
37:30
or research one, they tend to be more critical.
37:35
So we have, for example, what are called journal clubs where
37:39
we pick a part, a paper. And so we don't just absorb the
37:44
conclusions that the authors shared.
37:46
We talk about the limitations. We talk about, well, how could
37:50
this be wrong? But sometimes, you know, we
37:54
conclude, yeah, this is probably correct, but we're we're quite
37:58
critical, which, and again, being critical is a
38:03
fundamentally important part of good science, right?
38:06
Like we're all trying to get at the truth.
38:08
And the best way to the truth is just to critique.
38:10
And so also, as scientists, we want to be open.
38:14
To to critique. I'm only laughing because I feel
38:16
like maybe the truth is has become a subjective thing
38:20
instead of an objective thing, right?
38:23
That's that's. Yeah, that is a common, I think
38:27
that's the, the, the common human perception that we think
38:31
what we see is true. But and that's hence, you know,
38:35
again, that's the role of science is to recognize, note
38:38
what we see or what we think intuitively is correct is often
38:42
completely backwards. And hence we need experiments,
38:46
we need critique, we need a bunch of different perspectives
38:49
to to to contribute and discuss. I'm presuming here because you
38:55
brought up failure earlier, which is a question I wanted to
38:58
ask at some point. But based on what you just said,
39:02
what, how do you feel as a researcher or as a medical
39:08
trial, not trial participant, but trial conductor when the
39:13
results aren't what you're expecting, when you're, when
39:16
you, I don't want to say wrong, but when you say, hey, we have
39:19
this idea and we think this is what's going to happen.
39:21
Is there an emotional reaction that you have to that or is that
39:25
something you have to take away? How do you respond when things
39:28
don't go the way you expect? Yeah, so I have something I, I
39:32
saying, I repeat to my mentees, which is don't fight the data.
39:39
So we can be critical of the data.
39:42
We should double check. But it it is what it is.
39:48
And so yes, there is the sometimes the human emotional
39:54
response of disappointment, like, oh man, you know, that
39:58
didn't come out the way I wanted.
40:00
But often times it's the most exciting discoveries that are
40:07
counterintuitive and that are backwards.
40:09
And those are the things that are most novel, right?
40:13
If it's if, it works out precisely as we thought it would
40:16
at the beginning. That's not, that's not going to
40:19
be the most innovative, impactful thing.
40:22
And, and, and so it's something that we have to recognize in
40:28
ourselves. And again, I, I, you know, I try
40:31
to teach to my mentees, like embrace that when it doesn't
40:36
work out, you know, the, if there's a failure, that's an
40:39
opportunity to learn and grow. Another saying I have is strive
40:43
to be right, love to be wrong, meaning if you're wrong, you
40:49
know, of course, natural human response to that is to feel
40:53
badly. But if you look at it as like,
40:56
oh, that means I just learned something.
41:00
I just got closer to the truth. That's actually something to
41:03
celebrate in a weird, counterintuitive way.
41:07
So again, we're not trying to be wrong.
41:08
We strive to be right, but love to be wrong because that's an
41:11
opportunity to to learn and, and and and grow to to further
41:17
answer your question about, you know, the dissemination of
41:22
information not representing what we actually found.
41:26
So it is so I mentioned the medical audience.
41:28
Now, the lay media can be different and unfortunately,
41:33
and, and this again collides with human intuitions and human
41:37
emotions very clearly. You know, the media does not
41:42
like a complicated message. They want a single line 1
41:47
conclusion. And a good scientist, I'd like
41:52
to think, 'cause I, unfortunately, I, you know, I'm
41:54
not good at that. But I, you know, most scientists
41:57
should speak in terms of probabilities rather than
42:00
certainties. They shouldn't include caveats.
42:04
You know, things. The reality is complicated.
42:06
Reality is not that simple, and unfortunately that just doesn't
42:10
lend itself well to a compelling headline.
42:15
So yes, things are often oversimplified when it comes to
42:20
communicating to the play audience.
42:23
To that point, you were mentioning your mentees and
42:27
right before that you you brought up thinking critically
42:30
or examining data critically. How do you, how do we in this
42:35
current era teach people to think critically?
42:39
I feel like that's something that is, I don't want to say
42:44
look down upon, but it's just not the focus.
42:46
I think when I was, I'm, you know, I mentioned my age
42:48
already. So I think thinking, thinking
42:51
critically and problem solving was a part of my education.
42:56
Maybe I'm misunderstanding or misrepresenting the world to
42:59
simplify, simplifying things too much.
43:02
But we've turned things into more gamified or testing where
43:06
the answers are very specific and there's less critical
43:10
thinking. Is that are you seeing that in
43:11
the medical education that kids are getting?
43:15
Or how do you encourage your mentees to to do that critical
43:18
examination where they don't necessarily have an answer,
43:23
right or wrong? Well, yeah, I think, I mean, the
43:26
good news is that in the in the depths of academia and science,
43:33
I think that culture and that understanding of the importance
43:38
of being critical has been maintained successfully.
43:43
But unfortunately, I, you know, there are forces outside and,
43:48
and we have definitely recognized that there's a change
43:53
in the broader culture that probably gets in the way of some
43:58
of that. A concept that I found useful
44:01
and that I have shared with my younger mentees that they also I
44:05
think have found useful is the distinction between task
44:09
conflict and personal conflict. With the idea that task conflict
44:15
so taken the emotion out of it, the ego out of it, and focusing
44:20
on the problem at hand, the research question, the OR just
44:26
the question in general, task conflict is a good thing.
44:29
Generally, it's not a bad thing. Having discussion, having
44:32
debate, thinking carefully, being critical, those are all
44:35
actually productive things. They become counterproductive
44:40
when if someone takes them personally, so they're not able
44:44
to disentangle the task from the personal so and that goes in
44:48
both directions. So the supervisor or mentor
44:52
speaking to the supervisee mentee, they shouldn't be
44:56
attacking them personally, right?
44:58
Or an ad hominem sort of thing. They should focus on the work,
45:02
the task, the question. Similarly, though, the
45:06
supervisee, the mentee, if someone is criticizing their
45:11
idea, their product, their work, they shouldn't take that
45:15
personally. And again, we're all human.
45:19
That's not, I'm not saying that's easy, but this is the
45:22
goal. And especially if I think we're
45:24
explicit about that ahead of time and say, look, I'm going to
45:28
critique this. This is not a, you know, this is
45:32
not, this has nothing to do with your value, with my feelings for
45:36
you or your feelings. This is about the work.
45:38
And I'm trying to help you actually.
45:41
So again, I, you know, trying to encourage that sort of approach
45:46
and embracing that sort of critique, I think I mean, it's
45:51
more productive. It helps us arrive at the truth
45:54
and frankly, it helps with everyone's morale if we can kind
45:58
of have that shared understanding.
46:01
I had a coach when I was young who always told me if I'm not
46:05
yelling at you, it means I've stopped caring.
46:08
Now that's not necessarily great for a kid, but at the same time,
46:13
it's one of those things where if I'm not critiquing you, if
46:15
I'm not paying attention, it means that maybe the work that
46:18
you're doing isn't as effusive or valuable or something, But
46:23
the idea is that you're you're going back and forth with ideas.
46:26
I hope that I I didn't ruin that that thought.
46:28
I know that coming. We're coming up on 45 minutes
46:30
here in this chat and there are two things that I really wanted
46:34
to ask before. I know you've got a busy day and
46:37
then you get back. I certainly appreciate all this
46:38
time. The first one is, and just
46:42
because of who our audience is, coffee people, people interested
46:46
entrepreneurship, the industry at large, we're actually right
46:50
on the cusp of going to world of coffee.
46:52
So I want to get this this thought out there, but based on
46:56
the work that you've done examining the impacts of coffee
47:00
and or other beverages, what take away should should those of
47:03
us that that coffee is an integral part of our morning or
47:06
a part of our day. You know what take away from
47:10
these studies? Should we should would you like
47:13
people to have? Maybe not should we, but would
47:15
you say this is if I could? And I I'm doing that thing where
47:19
I'm simplifying. So I apologize for that.
47:23
But is there something that I should say?
47:24
Hey, it should keep in my mind or let you know marinate as we
47:28
go forward and I'm drinking my cup of coffee tomorrow.
47:31
Yeah. And maybe I I can briefly
47:33
summarize some of our other findings and then please answer
47:36
that question. I hope so.
47:38
Other findings in Crave that were important.
47:40
Again, these are secondary outcomes, but they were highly
47:43
statistically significant. So 1 was randomization of coffee
47:48
was associated with more step counts objectively determined by
47:52
the Fitbit, but less sleep. So on days when people were
47:57
randomly assigned a coffee, they had 30 minutes less sleep.
48:00
Another very interesting finding and, and this will all relate
48:04
back to the question you just asked is different people
48:08
metabolize caffeine at different speeds and we can actually
48:13
measure that by looking at their genes essentially.
48:17
So we got DNA from everyone and categorize them into slow,
48:22
intermediate and fast caffeine metabolizers based on their
48:25
genetics. And it turned out that the slow
48:30
metabolizers slept the least when exposed to coffee.
48:35
So they, they had almost an hour less sleep, whereas the fast
48:38
metabolizers had no difference in their sleep.
48:42
So one of the messages I would hope to communicate is 1 size
48:47
does not fit all when it comes to caffeinated coffee.
48:50
And you know, so the amount of caffeine that one is exposed to
48:54
could be the same between two different people, one drinking 1
48:58
cup a day and one having four to five cups per day.
49:01
So there's heterogeneity in terms of individuals and then
49:05
there's also heterogeneity in terms of good and bad.
49:08
So I would say having also published a review recently on
49:14
coffee specifically and various forms of heart disease, it is
49:18
quite remarkable, again, mainly based on the observational data,
49:22
how beneficial it appears caffeinated coffee is.
49:26
But it's not a panacea, it's for everyone.
49:29
It's not necessarily all good. And I would say this is the case
49:34
for almost every substance and I would include, you know,
49:36
moderate alcohol consumption here.
49:38
This that's a whole nother discussion.
49:40
It's not all good or all bad per SE.
49:43
There's it's again, this gets back to that headline versus
49:46
reality thing. It's more complicated than that.
49:49
And just to mention what we found in our decaf trial where
49:53
we were taking patients with this very common arrhythmia,
49:56
atrial fibrillation, people getting cardioverted, meaning we
50:00
were shocking them into a normal rhythm just as part of their
50:03
normal care, randomly assigned to consume at least One Cup of
50:08
caffeinated coffee a day versus avoid all coffee and caffeine
50:11
followed them for six months. The group randomly assigned to
50:15
caffeinated coffee had less recurrences of atrial
50:19
fibrillation. So providing quite compelling
50:22
evidence that caffeinated coffee seems to protect against this
50:28
common arrhythmia, atrial fibrillation at least among
50:31
people who previously consumed caffeinated coffee.
50:35
So again, there could it's complicated.
50:39
We cannot exclude the possibility.
50:40
There are some people went who consume coffee and feel their
50:44
heart racing. They're not going to, they will
50:46
not be in that study. So there could still be some
50:49
people who do have that effect, but for the great majority who
50:55
enjoy caffeinated coffee and, and who have this dangerous
50:58
arrhythmia, it actually reduces the risk.
51:01
So, you know, hopefully many interesting and, and and
51:06
hopefully useful lessons to take from these #1 conventional
51:11
wisdom is not always correct. And, you know, this is the
51:14
benefit of science and evidence. And following the evidence
51:18
number two, things tend to be pretty complicated #3 specific
51:24
to caffeinated coffee. The good news is, is that again,
51:27
the evidence suggests that it's quite tends to be healthy, tends
51:32
to be maybe have some cardiovascular benefits.
51:35
But keep in mind it's, it's not the same for everyone.
51:38
So you know, I wouldn't. So the way that I have
51:43
incorporated these most recent data into my own clinical
51:46
practice, for example, is to reassure people that enjoy
51:52
caffeinated coffee, that have not noticed a problem with it,
51:55
that it's fine to consume and having, you know, there's no
51:59
reason to stop because you're concerned that eventually you
52:02
might develop an abnormal heart rhythm.
52:05
However, for people that don't consume caffeinated coffee,
52:08
there's not sufficient evidence in my view to to encourage them
52:13
to start for the purpose of, you know, helping with their health.
52:18
Yeah, there's no one-size-fits-all and, and it's
52:21
why advocating for your own, your own medical health is so
52:26
important. And, and also being honest with
52:29
your, your medical professionals about what's going on in your
52:32
life. You can't hide some of the
52:34
information because you're embarrassed or you, you know,
52:38
you just, you got to really examine it and really get it out
52:41
there. And it's a relationship back and
52:43
forth. The last thing that I do on this
52:47
show. Well, not always the last thing,
52:49
but we've had, I've really enjoyed listening to, you know,
52:52
a lot of these thoughts. And I know that you can't really
52:54
break it all down in an hour. We're not trying to do that, but
52:57
get these, some of this out. I'm wondering if just on a more
53:00
personal level, we do this thing with roast our our presenting
53:03
sponsor, where we try to learn something valuable to our lives
53:08
in every show. I'm wondering, you do this job
53:11
that I feel could be overwhelming.
53:15
There's a lot going on. You're dealing with funding and
53:18
and government changes and the world is changing and the
53:22
perspective of medicine sometimes might be out there and
53:25
there's always things running around you.
53:27
How do you find balance in your life or in your mental state
53:30
when you're working in in a field like this that might be
53:33
overwhelming. Yeah, I think, you know, balance
53:36
is critical and Speaking of things being counter intuitive,
53:43
taking the time to do other things such as exercise and get
53:50
sufficient sleep can actually substantially enhance your
53:55
productivity. And so I again, this is
53:58
something I talk about with my mentees and something I try to
54:01
practice myself. You know, the best ideas I come
54:05
come up with are often, while I'm, I, I live in Marin County.
54:09
I'm very fortunate to be able to go on on mountain bike rides and
54:12
a beautiful Mount Tamilpius. It's often times when I'm
54:16
escaping some, you know, and really not deliberately thinking
54:21
about work that I come up with the best ideas and physical
54:27
activity, sleep, these kind of basic things, they are super
54:32
critical. And so you can spend all your
54:35
time working and not doing those things and you're not, you very
54:38
likely won't get as much done. So carving out the time and I
54:43
tell my patients, for example, especially when when I'm trying
54:46
to encourage them to exercise, make an appointment in your
54:49
calendar to exercise. I think that that helps a lot.
54:53
So it, it is absolutely critical to prioritize those things.
54:58
And you know, I don't want to speak just in terms of work
55:00
productivity. It also we need to live our
55:02
lives and enjoy them as much as we can.
55:08
And and that sort of balance, not to mention spending time, of
55:11
course, with family, with friends, with loved ones, is, is
55:15
just, is super important. I'm sure we missed a lot of
55:20
things, but is there one specific or are there a few
55:23
things, Any last thoughts that you might have that you really
55:25
wanted to get out to the Coffee People community today?
55:30
Well, I mean, I have to assume that the great majority of the
55:33
audience enjoys caffeinated coffee.
55:35
I, I share that enjoyment with you.
55:38
So, you know, I think especially in these difficult times where
55:42
we all, you know, where acrimony is, is all too common, hopefully
55:47
we can take a step back and appreciate that we have this
55:51
activity that we enjoy in common.
55:53
And, and it probably interestingly bridges a lot of
55:56
gaps. And as I've tried to communicate
56:00
throughout, you know, I, I can't emphasize enough how much a
56:04
science really does I, I feel contribute to the flourishing of
56:10
us humans. I mean, we really owe great debt
56:13
to it. It is super.
56:14
And for any young folks thinking about career choices, I, I
56:18
cannot encourage them enough to think about pursuing a career in
56:22
science. And that can take many, many
56:24
forms, but it really is fun. It's gratifying, it's
56:27
meaningful. And it and it and I, you know,
56:30
ultimately is, is the most helpful thing I think we can do.
56:36
Doctor Marcus, thank you so much for your time and for sharing
56:40
some of the information that we could glean today in an hour
56:44
with our show. And, and I hope that we get a
56:46
chance to have a cup of coffee again, maybe at some point where
56:48
we can see each other. Yeah, sounds good.
56:51
Thank you so much. It was fun.
56:52
You're welcome. The.