Feb 20, 2017
Carolyn Lam: Welcome to Circulation on the Run, your weekly podcast summary and backstage pass to the journal and its editors. I'm Dr. Carolyn Lam, associate editor from the National Heart Center and Duke National University of Singapore. I am so excited to be discussing the diabetic HFpEF or heart failure with a preserved ejection fraction phenotype, with world experts and new insights from the I-PRESERVE Trial. That will just be in a moment and here are your summaries first.
The first paper in this issue is a systematic review and meta-analysis of risk factors for Co-Arctation of the Aorta on pre-natal ultrasound. In this paper by first author Dr. Familiari and corresponding author Dr. D'Antonio and colleagues from Arctic University of Norway, the authors performed a systematic review of 12 studies on 922 fetuses with echo-cardiography, and found that those with a post-natal diagnosis of co arctation had significant differences in several cardiac morphological parameters compared to cases without co arctation. The presence of a co arctation shelf, or hypoplastic arch, was associated with a significantly increased risk of co arctation. Furthermore, they reported multi-parametric diagnostic models that were associated with an increased detection rate. Thus, this paper tells us that assessment of left inflow and outflow tracts prenatally may help in stratifying the risk of co arctation.
The next paper reports pre-clinical data that truly represents a paradigm shift in our understanding of vascular resident endothelial progenitors in tissue regeneration. In this paper by first author Dr. Patel, corresponding author Dr. [Cossroterrani 00:02:00], and authors from Royal Brisbane and Women's Hospital in Australia, the authors studied protein expression levels of common endothelial markers in mice using flow cytometry. They discovered an endovascular progenitor cell in vivo that is present in normal endothelium in the aorta and lungs and activated in vessel walls during various endogenic situations, such as in the placenta, skin wound healing, and tumors. They further define at a molecular level an entirely novel endothelial hierarchy from an endovascular progenitor cell to a mature different-shaded endothelial cell via complete RNA sequencing. They further clarify the linage of endothelial progenitors in their origin by using bone marrow transplantation and vascular-specific, lineage-tracing mouse models, showing that the endovascular progenitor cells were derived neither from bone marrow nor from hematopoietic progenitors. This discovery of an endovascular progenitor cell will have significant implications for the development of endothelial progenitors as a cell therapy.
The next paper addresses the chicken or egg question regarding the association between obesity and atrial fibrillation, and this is done using Mendelian randomization to define a causal association between body mass index and atrial fibrillation. In this study by Dr. Chatterjee and colleagues from Massachusetts General Hospital, the authors looked at more than 50,000 European individuals without atrial fibrillation at baseline and showed that genetic variance associated with increasing body mass index were significantly associated with an increased risk of atrial fibrillation. The association between genetically determined obesity and atrial fibrillation persisted even after adjustment for traditional atrial fibrillation risk factors, such as hypertension, diabetes, coronary artery disease, and heart failure. Taken together, these data are consistent with a causal association between increasing body mass index and incident atrial fibrillation. These findings therefore support the primordial prevention of obesity as a significant public health target to combat the expanding global burden of atrial fibrillation.
The last paper provides contemporary estimates of the stroke burden in China, a country which bears the biggest stroke burden in the world. In this paper by doctors Wang and Fagen from the Capital Medical University in Beijing, China and Auckland University of Technology in New Zealand, and colleagues, the authors reported results of a nationally represented door-to-door survey conducted in 2013 in 155 urban and rural centers in 31 provinces in China, totaling 480,687 adults. They found that the age standardized prevalence was 1,115 per 100,000 people, incidence rate was 247 per 100,000 person years, and mortality rates were 115 per 100,000 person years. The stroke prevalence estimates in 2013 were greater than those reported in China three decades ago, especially among the rural residents. Finally there was a north to south gradient of stroke in China, with the greatest burden observed in the northern and central regions. Well, that wraps it up for our summaries. Now for our discussion.
For our featured discussion today, we are talking about my favorite topic and of course that is HFpEF, or heart failure with preserved ejection fraction, and I am so thrilled to have with us today Dr. John McMurray from University of Glasgow, who's the corresponding author of our featured paper referring to diabetes in patients with HFpEF and really talking about the novel results from the I-PRESERVE Trial. Welcome, John!
John McMurray: Thank you Carolyn, it's always a pleasure to speak to you.
Carolyn Lam: Oh, I have been waiting for this one, and I'm so excited I don't know where to begin, but how about with this? Diabetes and HFpEF, first of all, haven't we spoken to death about co-morbidities in HFpEF? And secondly, what makes this paper special? Because we've heard about diabetes and HFpEF from CHARM, from DIG, from Relax, so tell us: why the interest in diabetes and HFpEF?
John McMurray: Sure, Carolyn. I think you and I have been interested in diabetes and heart failure, that terrible combination, for a long time. But I think there's a lot more interest in it today because, of course, we've had several new clinical trials with interventions to lower blood glucose that have showed both beneficial and potentially harmful effects on the development of heart failure. But really what these trials have highlighted is just how common heart failure is as a complication of diabetes. And we strongly suspect, though we don't know for sure of course, but we strongly suspect that most of that heart failure developing amongst patients with diabetes is probably heart failure with preserved ejection fractions. So, I think the context currently is that what's different about our study compared to the ones that you mentioned is that in I-PRESERVE we measured a number of things that were not available in, particularly, the large clinical trials previously. So, in I-PRESERVE we measured natriuretic peptides, we looked at health-related quality of life, and maybe most importantly of all we had a large echo-cardiographic sub study. So I-PRESERVE is quite different than DIG-Preserved and CHARM-Preserved, and of course a lot larger than the RELAX HFpEF study.
Carolyn Lam: I was the associate editor managing your paper and I was so excited about this that I invited an editorial as well by Brian Lindman, and he's got this beautiful table that summarizes what your study really adds to the literature, and I think it's so critical. Could you start by summarizing? What are the main findings?
John McMurray: Well, I-PRESERVE, as you know, was a trial of just over 4,000 individuals with HFpEF defined clinically and with an ejection fraction of 45% or above. There was actually a trial comparing the angiotensin receptor-blocker [inaudible 00:09:17] placebo, though in fact there is no difference in morbidity and mortality between those two treatment groups. So we've looked, as you said, at the patients who had diabetes and compared those to the patients who didn't have diabetes. I think there was some very interesting novel information; if you look at the two subsets of patients, they actually don't differ in terms of age and sex and, importantly, in left ventricular ejection fraction.
But there are other differences that you would expect; for example, many more of the patients with diabetes were obese. But interestingly, and despite that, the patients with diabetes had higher, significantly higher, NT-proBNP levels. So as you know, obesity tends to be associated with lower rather than higher natriuretic peptide levels, so here we were finding higher natriuretic peptide levels in a subset of patients who were actually, by and large, more obese. And there was no difference in other things that might have accounted for that difference; natriuretic peptides, for example, there was no difference in the proportion of patients who had atrial fibrillation.
So that was important, and that's also important when we come to think of outcomes because of course the previous studies reporting worse outcomes in patients with diabetes had not adjusted for natriuretic peptides because they by and large weren't available in the large prior trials. So that, of course, could have accounted for some of the worse outcome.
Some of the other things, features, maybe to pick out in terms of baseline characteristics ... one was that these patients had many more features of congestion, so patients with diabetes had more edema, more often had a raised jugular venous pressure and so on, and that's interesting given some of the recent clinical trial data that we might come back to. And even though the [inaudible 00:11:22] class distribution was not different between patients with diabetes and those without, what was very different was health-related quality of life, which was much worse in patients with diabetes than those without. Now you could if you chose to, Carolyn, look at that as saying that physicians weren't assessing worse functional status or symptomatic status in the patients with diabetes, but the patients were certainly self-reporting a much worse health-related quality of life.
So those were the, sort of, clinical characteristic differences. We did, as I said, have an echo-cardiographic sub study. There were 745 patients in total in the trial who had a detailed echo study, and there were perhaps more modest differences than I might have expected (and I'd be interested in your opinion about this) in patients with diabetes. So they had a somewhat greater, statistically significantly greater left ventricular mass, they had increased early diastolic mitral inflow velocity through E, they certainly had increased E over E prime increased left atrial areas, so there was some left ventricular remodeling and there was some evidence of increased left ventricular filling pressure, maybe diastolic disfunction. But the differences were not very striking; they were there, and as I said previously, ejection fraction (which most of us regard as perhaps not a very good measure of systolic function) was similar between the two groups. We didn't look at more sophisticated and [inaudible 00:13:09] measures of systolic functions so those could have been different, we just don't know.
So that's the baseline clinical features and baseline echo-cardiographic findings. And then, of course, we followed these patients for a median of just over four years and what we found was that the cardiovascular and all cause mortality was about twice as high in patients with diabetes as in those without. And if you adjust for conventional clinical variables, including NT-proBNP, which is individually the most powerful predictor of outcome, you only very slightly attenuate that greater risk associated with diabetes. The risk of heart failure and hospitalization was also about doubled in an unadjusted analysis, but that was more attenuated in the adjusted analysis. But you've also got to remember that, of course, the patients with diabetes were not surviving as long, so the very fact that they had a substantially higher risk of heart failure and hospitalization despite a shortened longevity is important.
Then lastly, again I think a fairly unique aspect of this study was that we then added the echo-cardiographic findings into the multi-variable model [inaudible 00:14:33] because it was only a subset of patients in which we had echo-cardiographic measurements. The statistical reliability of this is not as robust as in the main model, but what we saw was that there was more attenuation of the risk associated with diabetes when you added in the remodeling and diastolic dysfunction findings that we saw in the echo-cardiographic sub study. So that's a summary, I think, of the key points.
Carolyn Lam: John, I was really impressed and struck by the consistency of the message, which is what I really appreciated. What you added to the field was this consistent message that the diabetic HFpEF just had more signs of fluid overload in general, be it clinical, be it by NT-proBNP, be it by echo. And I thought that was something you said it was a moderate difference by echo; it was enough to be convincing to me, and I really appreciated that. The fact that adding the echo findings attenuated the significance ... you know we went back and forth about that quite a bit together, didn't we?
John McMurray: We did.
Carolyn Lam: I think at the end it is consistent, it is useful information. It tells me that perhaps some of these outcomes are mediated by this access fluid, to me, at least part of it. And I think that is how we ended up expressing it in the final paper.
John McMurray: I think you are absolutely spot on, Carolyn, because I don't think I had anticipated that the features of congestion would be so different. And you are correct in that, of course, correlates very well with natriuretic peptides, with the left atrial enlargement and so on.
And then of course (and this is clearly extrapolation) but then of course it makes one wonder about some of the trials with diabetes drugs that we've seen. The TZDs, glycosomes, which calls a little bit of fluid retention, of course precipitating heart failure, and then the opposite recently with the SGLT2 inhibitors which of course are diuretics, and those drugs preventing the development of heart failure.
And it does make me wonder if the diabetic phenotype maybe was a little bit of renal dysfunction, some subtle renal dysfunction, is a sodium and water avid phenotype state and that it doesn't take very much to tip those patients into frank heart failure and perhaps we need to think (and I think you might have been alluding to it) think about insuring that we adequately diurese these patients given that in this study where people were supposed to be optimally treated, clearly there was still a lot of evidence of residual fluid overload.
Carolyn Lam: I absolutely agree, and yes you read my mind that I was going to allude to the implications for therapies that have a diuretic effect, you know, like the SGLP2 inhibitors and in fact this was discussed in Brian Lindman's editorial, which is a must read.
Just another question though. What do you think of peripheral mechanisms contributing to all this?
John McMurray: Yeah, obviously there is the kidney aspect that we saw a relatively small difference in estimated GFR. Of course that only tells you one aspect of renal function and the nephron in diabetes may well be sodium avid maybe more likely to retain water. So certainly the kidney as a peripheral mechanism might be very important.
And then of course the blood vessels, I mean there's no question that patients with diabetes have more abnormal endothelial function probably have got enhanced vascular stiffness. And of course we know from a long time ago at least in HFrEF (I'm not so sure about HFpEF) but in HFrEF there's evidence that some of the vascular stiffness you see in patients with HFrEF is actually due to sodium in the vessel wall and there's some beautiful old-style clinical physiology experiments showing that if you diurese patients with HFrEF you restore vasodilation you restore basal motor responsiveness. It could also be true in HFpEF though of course patients with HFpEF and many other reasons to have vascular stiffness.
So yes, peripheral mechanisms may well be important. Your humoral abnormalities may be more pronounced in patients with HFpEF and diabetes compared to those without diabetes. We don't know because I'm not sure that's been measured very often. Certainly natriuretic peptides are, but what about things like the angiotensin system and arginine/vasopressin and the sympathetic nervous system. You know, there's still so much to study looking at patients with heart failure with and without diabetes because they're really quite distinct. And whatever's going on it makes a big difference the way those patients feel, what they can do, and what happens to them.
Carolyn Lam: Yeah, and your study really establishes that. Congratulations once again John, it's just been such a delight chatting with you.
John McMurray: Likewise, Carolyn.
Carolyn Lam: Listeners, you heard it right here on Circulation on the Run. Don't forget to tell all your friends about this podcast and turn in next week!