Jul 14, 2016
Dr. Lam: Welcome to Circulation on the Run. Your weekly podcast
summary and backstage pass to the Journal and it's editors. I'm Dr.
Carolyn Lam, Associate Editor from the National Heart Center, and
Duke National University of Singapore.
Dr. Sanjay Kaul and Darren McGuire will be joining me in just a
moment to share their perspectives on the EMPA-REG OUTCOME trials.
Are the results with empagliflozin in diabetic patients at high
risk, are they too good to be true. First, here are the highlights
from five original papers in this week's issue.
The first paper is from Dr. Gilboa, from the National Center on
Birth Defects and Developmental Disabilities, Centers for Disease
Control and Prevention in Atlanta, Georgia, and Dr. Marelli from
the McGill Adult Unit for congenital heart diseases in Montreal,
Quebec, and colleagues. These authors recognize that because of
advancements in care there has been a decline in mortality from
congenital heart defects over the last several decades. However,
there are still no current empirical data documenting the number of
people living with congenital heart defects in the United
States.
These authors address this gap in knowledge by using prevalence
data from Quebec, Canada, in the year 2010, as a foundation for a
mathematical model, and estimated that in the United States in the
year 2010, approximately 2.4 million people, including 1.4 million
adults, and 1 million children were living with congenital heart
defects. This estimate is significant, because it corresponds to a
63% increase in the estimated size of the adult population with
congenital heart defects in the United States since the year 2000.
This has significant implications for resource allocation for
health services delivery that will need to account for this growing
population of adults with congenital heart defects.
The second paper is from first author Dr. Tabot, and corresponding
author Dr. Liao, from the University of Chicago, and colleagues who
aim to understand better the common complication of angiodysplasia
leading to nonsurgical bleeding in patients with left ventricular
assist devices. The authors studied 101 patients with heart
failure, left ventricular assist devices, or orthotopic heart
transplants. They found that compared to patients with heart
failure, or transplant patients, patients with left ventricular
assist devices had elevated serum levels, and endothelial
expression of angiopoietin-2, which is a potent angiogenic
mediator.
Elevated levels of angiopoietin-2 in these patients increase
angiogenesis in vitro, and were associated with bleeding events.
Furthermore, they found that increased thrombin levels in left
ventricular assist device patients were associated with elevated
angiopoietin-2 levels. In aggregate, therefore, the results
indicate that high levels of thrombin induced endothelial
angiopoietin-2 expression, which may then contribute to
angiodysplasia and non-surgical bleeding in patients with left
ventricular assist devices. The clinical implications are that
clinical studies angiopoietin-2, and factor 12 inhibitors may
therefore be indicated to prevent nonsurgical bleeding in patients
with left ventricular assist devices.
The third paper is Dr. Gordon from Hasbro Children's Hospital in
Rhode Island, and Dr. Kieran from the Dana Farber Cancer Institute
in Boston, Massachusetts, and colleagues who addressed the
Hutchinson Gilford Progeria Syndrome. An extremely rare, fatal
segmental premature aging syndrome, where without specific
treatment, death usually occurs at an average age of 14 1/2 years
from an accelerated atherosclerosis.
A PRIA single arm clinical trial has demonstrated that the protein
farnesyltransferase inhibitor, Lonafarnib, ameliorates some aspects
of cardiovascular and bone disease in this syndrome. The current
trial sought to further disease outcomes by additionally inhibiting
progerin prenylation using pravastatin and zoledronic acid on top
of Lonafarnib in 37 participants with the Progeria syndrome.
Results showed that the composite primary study outcome of
increased rate weight gain and decreased carotid artery echodensity
was achieved. Overall, participants experienced increased bone
density, size, and structural properties. However, unlike the PRIA
single arm Lonafarnib monotherapy trial, mean carotid-femoral pulse
wave velocity and mean carotid artery adventitial echodensity were
not improved. In addition, rates of carotid and femoral artery
plaques and extraskeletal calcifications all increased.
In summary, compared PRIA Lonafarnib monotherapy treatment,
additional bone mineral density benefit, but likely no additional
additional cardiovascular benefit was obtained with the addition of
pravastatin and zoledronic acid. The authors concluded that since
increased bone fracture is not a disease feature, the addition of a
combination of statin and biphosphonate to Lonafarnib therapy is
not recommended for clinical treatment of Progeria syndrome.
However, it is reasonable to consider statins if concurrent lipid
abnormalities need to be treated.
This paper is accompanied by an excellent editorial by Dr. Francis
Collins, who describes our journey in seeking a cure for this rare
disease of Progeria.
The fourth paper is by first author, Dr. Grisenti and corresponding
author Dr. Tilley from Lewis Katz School of Medicine, Temple
University in Philadelphia, and colleagues who aimed to better
understand the role of leukocyte expressed beta-2 adrenergic
receptors in regulating immune cell responses to acute cardiac
injury. The authors achieved this aim by studying wild type mice
who were irradiated, and then transplanted either with isoform
specific beta adrenergic receptor knock out bone marrow, or wild
type bone marrow. These chimeric mice, after full reconstitution
then underwent myocardial infarction surgery.
Results showed that immune cell specific beta-2 adrenergic receptor
expression was essential to the repair process following myocardial
infarction. In the absence of beta-2 adrenergic receptors, vascular
cell adhesion molecule-1 expression was increased in leukocytes,
inducing their splenic retention following injury, and leading to
impaired scar formation, followed by rupture and death. Splenectomy
partially restored the beta-2 adrenergic receptor deficient
leukocyte infiltration into the heart, but gene therapy to rescue
the leukocyte beta-2 adrenergic receptor expression completely
restored all injury responses back to normality.
This study is clinically important because it highlights a bit of a
tension that we're facing. On the one had, beta adrenergic
receptors are known to regulate cardiac function and remodeling
following myocardial injury, by their effects through
cardiomyocytes. That's why we use beta blockers to prevent, at
first, cardiac remodeling. However, the current studies now
indicate that inhibition or deletion of the immune cell expressed
beta-2 adrenergic receptor causes leukocyte dysfunction, and
impaired immunomodulatory responses to myocardial injury.
These results may, therefore, have implications on the use of beta
blockers around the time of acute myocardial injury, such as
myocardial infarction, or perioperatively. This is really an area
that needs further research and understanding.
The fifth paper is by Dr. Herman, from the hospital of the
University of Pennsylvania, and colleagues who report on the one
year clinical outcomes of SAPIEN 3 transcatheter aortic valve
replacement in high risk and inoperable patients with severe aortic
stenosis. Now, as a refresher, in the initial partner trial of
transcatheter aortic valve replacement for high risk and inoperable
patients with severe symptomatic aortic stenosis, there was a
demonstration of marked survival advantage compared to medical
management ... But a high one year mortality of 24% in the high
risk, and 31% in inoperable patients.
More recently, the lower profile SAPIEN 3 prosthesis system has
become available. Which has a balloon expandable cobalt chromium
frame, with bovine pericardial leaflets, and an external fabric
seal. The early 30 day outcomes of this system have been reported,
and show a very low rate of adverse events.
The current study now reports the one year survival, and showed
that all cause survival was more than 85% for all patients, above
87% in the high risk, and above 82% in the inoperable subgroups.
Furthermore, there was a high rate of transfemoral access at 84%,
and a high all cause and cardiovascular one year survival in the
high risk transfemoral subgroup of 89% and 93%, respectively.
Between 30 and 365 days, the incidence of moderate perivalvular
aortic regurgitation did not increase. There was no association
between mild perivalvular leak and one year mortality. Although, a
small increase in disabling stroke occurred.
These results, which likely reflect device iteration and procedural
evolution, support the use of Taver as a therapy to consider in
high risk and inoperable patients with aortic stenosis.
Those were the highlights from this week's issues, and now for our
feature paper. We will be discussing the perspective paper entitled
"Is the Mortality Benefit With Empagliflozin in Type 2 Diabetes Too
Good to be True?". To discuss this, we have two very special
guests. First, Dr. Sanjay Kaul, writer of this paper, and from
Cedars-Sinai Medical Center. Second, Dr. Darren McGuire, deputy
editor of circulation from UT Southwestern. Welcome, Sanjay and
Darren.
Dr. McGuire: Thanks, Carolyn.
Dr. Kaul: Thank you, Carolyn.
Dr. Lam: To start us off, I'd really love if Darren could please
introduce this new content category of circulation. Frame of
reference section, of which this is one of the papers, a
perspective article.
Dr. McGuire: Sure, so we envisioned, as we're evolving circulation
to our new editorship, an opportunity for authors, luminaries in
the field, to give us in a very encapsulated form, a laser focus
perspective on a specific topic. These come in two flavors, the
perspectives piece, which this is, is a little more evidence and
scientific quantitatively based. Then we'll also have a section
called on my mind, which is more of a free-flowing opinion
editorial targeting possibly a contentious or controversial issue.
These are going to be very short, and hopefully very entertaining,
and kind of teasers for the readership of the Journal.
Dr. Lam: Sanjay, you made it very personal, and I like that, too.
Share with us how this idea came about.
Dr. Kaul: Well, I was very impressed at the reception that the
results of the EMPA-REG outcome trial received at the EAST meeting
at Starcom last year. While I was witnessing the applause, I had
polar reactions. On one hand, I thought that after nearly five
decades of trials with checkered history, with regards to
cardiovascular outcomes, here we have for the first time a trial
demonstrating not only cardiovascular benefit, but a mortality
benefit. I thought maybe it's time to take the trumpets out and
sort of herald this holy grail, which we had failed to achieve. On
the other hand, realizing that we had been fooled before many times
by trials, yielding implausibly large treatments actually, that
were never replicated at subsequent trials.
I had a skeptical response to it, and sort of asked this question
rather tongue-in-cheek, or maybe used as a rhetorical tool to
address whether this mortality benefit was too good to be true.
Dr. Lam: You know, you didn't just question it. You examined the
data, and provided even more evidence. That's what I was impressed
with in your paper. That table where you provided base factor, as
well as a Bayesian analysis. Could you break that down for us, and
explain what you found?
Dr. Kaul: Yes, I was trying to sort of examine the strength of the
evidence, in terms of the quantitative aspect. Yes, the effect size
for the cardiovascular benefit was quite impressive. For the
primary endpoint, which was a compositive cardiovascular, death,
non-fatal MI, and non-fatal stroke, the p-value was not very
robust. It was .04. The p-value tends to overestimate the strength
of evidence. I utilized base factor, which basically is a metric
that allows the two competing hypotheses to predict the data. Using
the base factor, I was able to demonstrate that the alternative
hypothesis was stronger than the null hypothesis by eight-fold. The
p-value of .04 translated into a base factor of .13. Which is not
strong evidence against the null hypothesis. It requires
independent confirmation and subsequent trials.
A p-value of .04, while meeting the superiority criteria, would not
be sufficient enough to meet the FDA's requirement of substantial
effectiveness. Substantial effectiveness just basically means that
the FDA requires two trials, each with a p-value less than .05. In
1998, they modified their regulatory requirement, and accepted that
one single trial would be sufficient, provided that there would be
a persuasive p-value. Persuasive basically is defined as a p-value
less than .001.
The base factor allows us to sort of interpret the strength of the
evidence, with respect to the primary composite endpoint was not
strong enough to meet this requirement. With respect to
cardiovascular mortality, as well as all cause mortality, which
trumps all other endpoints, it was persuasive enough.
Dr. Lam: What's your conclusion on that?
Dr. Kaul: What is controversial about that was that in the three
specified statistical plan, the so-called hierarchical testing
strategy, the non-inferiority for three point MACE, followed by
non-inferiority for four point MACE, and followed by superiority of
three point MACE, and lastly, superiority of four point MACE.
Because the p-value of four point MACE superiority was .08, one can
argue purely from a statistical perspective that you stop your
testing strategy, and any analysis beyond that would be deemed
exploratory. Even though cardiovascular mortality and all cause
mortality was prespecified, the purist would argue that since you
failed superiority for four point MACE, you really can't proceed
further. You can analyze, but it will be considered an exploratory
analysis.
I sort of wept and said that because Christopher Columbus had
prespecified that he will be discovering the route to India, the
fact that he stumbled upon America does not mean it doesn't exist
because he had not prespecified it. I think all cause mortality is
the most meaningful endpoint, and the least subjective measurement
error. It meets the key attributes of regulatory decision making.
Which it's prespecified, it's highly persuasive, therefore, it
meets the replication criteria, and the p-value is so robust that
even if you adjust for nearly 100 multiple comparisons, the p-value
would still hold. It meets all the regulatory criteria for
approval.
Dr. McGuire: Sanjay, let me just chime in here. I think it's also
important, not only were these prespecified, but it's important, I
think, for readers of these diabetes programs to realize that
hospitalization for heart failure ... Although it's not part of the
primary outcome ... In virtually every one of these trials, it is
prospectively collected, chartered to find, and essentially
adjudicated by blind endpoint adjudicators. You know, death is
death. Cardiovascular death in these programs are all adjudicated,
as well. I think the prospective collection and central
adjudication also adds legitimacy to the hospitalization for heart
failure are above and beyond the analytic issues.
Dr. Lam: Darren and Sanjay, I hear both of you kind of saying it
does look like, even looking at it from different angles, the data
do look strong. At the end of the day, Sanjay, you concluded that
it does need another trial. Results do need to be replicated. That
was your conclusion. I'd love to hear Darren's take on this.
Dr. McGuire: I think what Sanjay is saying there, and I think what
we all believe, was we would really love to see this observation
with another member or members of the class. We're learning a lot
in hindsight based on these observations, and people are exploring
potential mechanistic underpinnings. We're learning a lot about the
mechanisms of these medications, above and beyond their glucose
uric effects. There's a lot of implication about renal physiology
and hemodynamics, and altered myocardial metabolism. I think as
Sanjay points out in the paper, some of this looks like a possible
arrhythmic effect. We have a lot to learn about this mechanism of
action, and whether or not this will be unique to impact
gliflozin.
It has been publicly announced, Boehringer Ingelheim is planning,
they're in the planning phases for heart failure trials with
empagliflozin to further explore this signal. I think they will
address Sanjay's desire to have some replication in a different
patient population. Still, we would love to see these extended into
other patient populations. To both extend the use of the
medications if they're found, but also provide further confirmation
of the observations from EMPA-REG outcome.
Dr. Kaul: Carolyn, let me also add, I used the title as a
rhetorical tool, as I stated earlier. I do conclude that the
mortality data is not likely to be spurious. In the back of my
mind, I still have that 1% skepticism that I would like to
eliminate, because the findings were totally unexpected, and
unprecedented, as we discussed earlier. If all the pathways,
including the mechanistic pathways are aligned, I would have
substantial reassurance, beyond any reasonable doubt that the
findings are true. That's why I'm asking for replication. Not
necessarily by empagliflozin in other trials, but by another
molecule within the same class. I think that would be
sufficient.
Dr. McGuire: Yeah, and I think it's really interesting to note
there, is that I was involved in the early days of some of these
drugs as they're being developed. When the other two members of
this class went to the FDA, dapagliflozin and canagliflozin, they
provided FDA's requirement and meta analysis from all of the phase
2B and 3 trials that had been completed to date. The meta analysis
of the cardiovascular outcomes. Both dapagliflozin and
canagliflozin had point estimates of cardiovascular death reduction
of 30%, and 35%, respectively. When we saw those data, they were
based on 25 to 40 total events. We chuckled, thinking this is
spurious, from small events being analyzed. That there's no way
they would prevent cardiovascular death. Sure enough, you know, you
could almost superimpose those point estimate plots from the phase
2B-3 meta analysis, with the ultimate outcomes from EMPA-REG.
There's some promising, although again, very statistically
imprecise estimates that this may well be a class effect. As many
of the listeners will know, there are ongoing cardiovascular
outcomes trials for all of these medications. That will come some
time in the next year or two.
Dr. Lam: That's fantastic. Thank you both for sharing those
perspectives. I mean, I learned so much. I really think, Sanjay,
your paper achieved exactly what you had meant for it to achieve,
and exactly what circulation was hoping to create the discussion,
as well.
Dr. McGuire: Thank you, Carolyn.
Dr. Kaul: Thank you very much.
Dr. Lam: You've been listening to Circulation on the Run. Thank you
for listening. Don't forget to join us next week for more
highlights and discussions.