Circulation April 17, 2018 Issue

Dr 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. Does NT-proBNP-guided therapy improve outcomes in acute decompensated heart failure? Well the Prima II trial results are coming right up after these summaries.

                                Is hospital volume a good structural metric assessing the quality of care in heart failure? Well, in the first original paper this week from Dr. Kumbhani and colleagues at UT Southwestern Medical Center, authors determined the relationship between admission volume, process of care metrics, and short and long-term outcomes admitted with acute heart failure in the Get With the Guidelines-Heart Failure registry, which has linked Medicare in patient data at 342 hospitals.

                                They found that lower volume hospitals had worse adherence to important heart failure process measures, than higher volume hospitals. There was no association between risk adjusted in-hospital mortality and hospital heart failure admission volume among older adults.

                                After adjusting for adherence with process measures at discharge, annual heart failure admission volume had a minimal association with mortality, and readmissions up to six months post-discharge. Thus, rather than focusing solely on hospital volume, hospital profiling efforts should perhaps focus more on participation in quality improvement initiatives, adherence to process metrics, and risk standardized outcomes.

                                The next study describes the association between air pollution and heart disease mortality in the United States, with a focus on whether the association differs by race and ethnicity. First and corresponding author Dr. Jennifer Parker from the National Center of Health Statistics Centers of Disease Control and Prevention and her colleagues use data from the 1997 to 2009 National Health Interview Survey linked to mortality records through December 2011 and the Annual Estimates of Fine Particulate Matter or PM2.5 as an index of air pollution.

                                They found that the association between air pollution and heart disease mortality in this national sample was elevated and similar to estimates found in prior studies. After controlling for social demographic and geographic factors, the associations between air pollution and heart disease mortality for non-Hispanic black and Hispanic adults were not statistically significantly different from that of non-Hispanic white adults.

                                Thus, this study supports the application of findings from prior studies of air pollution and mortality, albeit largely from non-Hispanic white adults, but to other races and ethnicities in the United States.

                                The next study suggests that large cardiac muscle patches engineered from human induced pluripotent stem cells may be a reality. First author Dr. Gao, corresponding author Dr. Zhang from University of Alabama at Birmingham generated human cardiac muscle patches of clinically relevant dimensions of 4 x 2 centimeters and they did that by suspending cardiomyocytes, smooth muscle cells, and endothelial cells that had been differentiated from human-induced pluripotent stem cells in a fibrin matrix and culturing this construct on a dynamic platform.

                                The results from in vitro assessments of calcium transience, action potential propagation, and forced generation, as well as the presence of intercalated disc-like structures, suggested that cardiomyocytes matured in these human cardiac muscle patches. During the 7-day dynamic culture period. When transplanted onto infarcted swine heart, measurements of cardiac function, infarct size, wall stress all improved with no increase in arrhythmias.

                                Changes in the expression profile of myocardial proteins indicated that the human cardiac muscle patch transplantation partially reversed abnormalities in sarcomeric protein phosphorylation. Collectively, these observations indicate that human cardiac muscle patches can be successfully generated and may improve recovery from ischemic myocardial injury.

                                Does a second arterial conduit improve outcomes after multivessel coronary artery bypass grafting? Well, in the next study from first author Dr. Goldstone, corresponding author Dr. Woo, from Stanford University and their colleagues used a clinical registry including all 126 non-federal hospitals in California to compare all-cause mortality, and rates of stroke, myocardial infarction, repeat revascularization, and sternal wound infection between propensity score matched cohorts, who underwent primary isolated multivessel coronary artery bypass grafting with the left internal thoracic artery, and who received a second arterial conduit or a venous conduit between 2006 and 2011.

                                The authors found that receipt of a second arterial conduit was associated with lower mortality, and at first cardiovascular events, compared with receipt of a venous conduit. The survival benefit associated with the use of a second arterial conduit extended to patients up to 78 years old. As a second arterial conduit, the right internal thoracic artery offered no benefit, compared with the radial artery, but it was associated with an increased risk of sternal wound infection.

                                These findings therefore suggest that surgeons should perhaps consider lowering their threshold for using arterial grafts and that the radial artery may be the preferred second conduit.

                                That wraps it up for our summaries. Now for our future discussion.

                                NT-proBNP and natriuretic peptides in general, have really become mainstay in management of heart failure, in the diagnosis, in the prognostication, but questions still remain regarding NT-proBNP-guided therapy. We heard about the guided trial in chronic heart failure just reported last year, and this year, in fact this week, in this week's journal, we're about to hear about PRIMA II trial in acute heart failure.

                                And how NT-proBNP was tested as a potentially guiding strategy for the management of acute heart failure. I'm so pleased to have the corresponding author with us, Dr. Wouter Kok, from University of Amsterdam, as well as our Senior Editor Dr. Biykem Bozkurt from Baylor College of Medicine. So welcome both of you, and Wouter may I just jump straight in it?

                                PRIMA II means that there was a PRIMA I trial, so could you just briefly tell us a bit about PRIMA I and the rationale for PRIMA II?

Dr Wouter Kok: Well the PRIMA II was an in-hospital guiding therapy that was preceded by the PRIMA II, it was a chronic heart failure patient population and one of things that we noticed in PRIMA I was the lack of effect of trying to reach a percentage drop in chronic heart failure patients. Why is that? Is that because there is a long time before you can achieve a therapy adjustment? Or is it something else? And shouldn't we start before patients are discharged from hospitals?

                                So the idea was born to do an in-hospital guiding study instead of chronic heart failure patients study.

Dr Carolyn Lam:                Interesting. And could you tell us briefly, the design of PRIMA II and your findings?

Dr Wouter Kok: So the PRIMA II was designed based on the previous publication of several authors indicating that a 30% reduction in NT-proBNP would be a good target for heart failure therapy. Now, we first asked ourselves the questions, whether we should put this target in front of the hospital admission, so in the first 2 days or perhaps at the end of the hospital admission? And the 30% reduction was validated only for discharge purposes, so but we also tried to establish whether we could precede this date a little bit before discharge, but it appears that you cannot precede it too much.

                                So you cannot do it at day 3 or day 4, when patients are not stable. Because then you may expect a rise in proBNP again before discharge, and then you already ran the rise patients to discharge. So we decided to do it at discharge. At least 1 or 2 days before discharge, when patients would be clinically stable. And this definition of clinical stability was important because there should be one guideline for doctors to say, OK this patient has been treated well, or not.

Dr Carolyn Lam:                Interesting. And so patients were randomized only after clinical stabilization, though in hospital after an acute decompensation, right? And then maybe the randomization arms, and the results please?

Dr Wouter Kok: Yeah, so the patients were randomized about day 7 or 8 after clinical stabilization, and day 3, but also patients at day 9, but when they were stable, they were randomized. And then the proBNP was measured, and when it was not reduced more than 30% they were guided. And when they were reduced more than 30% they were not guided but they were made ready for discharge.

                                So this was the randomization group. And the conventional group, the NT-proBNPs were measured at the randomization, and also at discharge, but nothing was revealed to the doctors. So it was only as a comparison for example, in the number of days necessary to wait before discharge, if this would influence the results.

                                The main finding is that the end point was negative for total mortality after 6 months, in combination with heart failure readmissions. So there were about 36% end point in both groups.

Dr Carolyn Lam:                Yeah Wouter, you know, we've just come from the guided trial that was so soon neutral and infect, ended early and that was in the chronic heart failure setting so very different from what you tested in PRIMA II. Congratulations first of all for a beautifully done study.

                                But may I just ask, because in guided it was mentioned repeatedly that perhaps even the control arm was treated so well because these were such specialized centers. So what kind of centers took part in PRIMA II?

Dr Wouter Kok: We started at centers in Amsterdam, they were all very well educated in heart failure treatments, and all were using proBNP before the study started, so they were experienced in interpreting proBNPs. Because we had too little centers, and the inclusion rate was not so fast, then we asked other centers to participate, and we asked 2 for instance in Barcelona and Porto in Portugal, which helped us to complete the trial.

Dr Carolyn Lam:                Oh that's really nice. And the design is really quite special, and I'm so appreciative that you took the time to explain that they were randomized only when stabilized.

                                Biykem, what do you think of that?

Dr Biykem Bozkurt:         It's a fascinating trial, I have to congratulate Wouter and his colleagues. The number one very important finding I think is, about two-thirds of the patients before randomization are able to achieve reduction of NT-proBNP more than 30%. So subsequent to that in the guided therapy we're able to achieve maybe an incremental additional 15% adding to about, I think 80% of the patients initially randomized to the NT-proBNP arm. Achieving a reduction more than 30%. So overall, if the patient's naturally before randomization, achieve a reduction NT-proBNP, two-thirds of the time, pushing it further, trying to achieve a further dry state, by randomization does not appear to make any changes in readmission rates, or mortality at six months.

                                So this very important finding is the majority of the patients on conventional strategies are able to be decongested and achieve clinical stability. Now the other important finding is, I think about 17-20% of the patients regardless of what we do, do not demonstrate this significant drop in their NT-proBNP levels. Which I call as a non-responder team, which is a fascinating group of individuals. So we have the yin-yang, individuals may actually demonstrate that they're responsive. And when they're responsive, then the majority of the patients do demonstrate a reduction by more than 30%, and even if we push it further by targeted therapies, don't make a difference in outcomes.

                                About 17-20% regardless of what we do, do not respond, and from former studies we know that those patients are associated with worse outcomes. The other important finding I think, is what changed in the study? What medications, what therapies were changed? It was fascinating from Wouter's group to recognize that there was a little, significant, but a little increase in the ACE admission prescription. But there was also an interesting finding in the guided therapy, that the beta blocker used was slightly lower.

                                That raises a question of if we were to just chase the numbers, meaning try to just target therapies according to the NT-proBNP levels, whether we would see some unintended consequences such as reduction in medications, just because the numbers may be going in one way or the other. This is acknowledged in Lynne Warner Stevenson's editorial that will be accompanying the paper. And the editorial is very nicely titled "Getting to Dry". So I found that fascinating to recognize that the therapies, when especially the conventional arm is treated well, did not differ.

                                As was the case in the guided trials. When you treat the patients very well, as was seen in this trial, there was not much of a difference. But again trying to treat a number by targeted therapies may not result in all the optimization that as we envisioned to see. And the third concept is the length of stay, of course in the U.S. is a major issue, and I do realize when we're trying to treat a number, sometimes the length of stay may end up being longer. And I do realize that perhaps in the targeted therapy group, the length of stay was a little bit longer, maybe Wouter can comment on that.

                                But overall it didn't result in any change in outcomes, or was not associated with any of the outcomes. So that was also an interesting finding. Because we tend to focus a lot on length of stay, but interestingly I guess by secondary analysis, there was no association with the clinical outcome.

Dr Carolyn Lam:                Wouter, would you want to comment the length of stay concept?

Dr Wouter Kok: Well it's indeed in the guided group, and the randomized group who were trying to attain the 30% NT-proBNP reduction, the length of stay was longer. Something about 11 days, compared to those who did not need guiding was about 8 days. Still long compared to U.S. standards, but it was the same in the conventional group, so about 9 days is respective of whether they reached a 30% reduction or not.

                                So here is the clinical experience. So the patient cannot tell whether he is reduced more than 30%, and the doctor isn't able to tell either. Because then the admission would have been longer probably. But trying to lower the 30% more, has some effect. There's little effect, but it has some effect. And then they have to do a sort of economic analysis, is 3 days longer in hospital, is it worthwhile to do that compared to for example reduction in admissions that you receive? This is a small population, only one-third of the patients who need guiding, and more than half of them you will reach somewhat more reduction than if you don't try at all.

                                So for us, that is the main result of the trial, if there is a signal, then it is still possible to do something, and the other remark about whether you increase or decrease medication, that's something that was discussed in the guided study too. So what is the best for the patient, is that the maximum medication or not, and we see for example, that if we reduce beta blockers, in some patients then some will improve in their functioning and also in the BNP.

                                So it's not always necessary to increase and increase medication. So that was also some signal that we tried to do some more research in. What is the target? Is the target a guideline, saying that more medication is better? Or is the target itself for proBNP a possibly better target than that?

Dr Biykem Bozkurt:         And the other interesting finding for that, there were no differences in chemo concentration levels in the guided versus non-guided groups. And last point that I wanted to make is the larger BNP reduction was amongst the individuals who did not require any guidance in successfully guided versus unsuccessfully guided, compared to those who did not need the guidance.

                                Those who were able to achieve the more than 30%, when you look at the magnitude, meaning amongst the individuals who are going to naturally respond to therapy, the natural responders, the decrement, or the decrease in the BNP levels are larger, than those ones we're trying to push. So that was another interesting, fascinating ... I was almost thinking whether that in the future we should look at responsiveness of patients, if we see they're responders then try to target their therapy or not.

                                So in a sense the non-responders, they now respond regardless of what we do. Responders may be gaugeable or titratable, or maybe with the precision respond to targeted therapies that almost have a dichotomous approach. What do you think about that Wouter?

Dr Wouter Kok: I say yeah we made a big mistake in thinking that more than 30% for patients who still needed guiding would be the same as rating the more than 30% without guiding. But the difficulty you have in reaching the 30% is already indicative somewhat less increase in prognosis than you will reach it spontaneously.

                                So we have to adapt our numbers for the trial, so it is recalculation that how many patients we would need to be successful in our trial, and that would be 600 patients in every arm, and then even then, you have to recalculate some of the effects that you will have to reach them. Perhaps the mid-range risk group is a better risk group to target than the highest risk group. That's something that we have to think about too.

Dr Biykem Bozkurt:         I think we will probably need to focus on individualization, I almost feel as though we will need to learn from the cancer trials, and see whether we could try to target rather than you know the population based clinical trials, trying to do the targeted therapies. Maybe fine tune the ability to precisely target, and of course that requires a little bit more layering of the markers and or a signal that we're going to be profiling in the individual.

                                So I don't think it's the end of targeted therapies, perhaps requiring a little bit of a more precision, and maybe individualization. But I am fascinated by first realizing it's a responder, and then maybe trying to accelerate and or optimize therapy, perhaps especially when we are forced or driven by administrative concepts such as length of stay or others. So making sure that maybe these variables, these biomarkers may help us recognize that maybe we haven't achieved that appropriately dry state yet.

                                But those all need to be determined, of course, by future trials, so far targeted therapies both in the acute and in the chronic does not seem to result in implementing outcomes.

Dr Wouter Kok: Well and the next step for us is to try and think how can reduce proBNP in all patients, we tried it with medication, but didn't do that much of catheterizations for those who were ... there were 50% of patients who were ischemic so why don't we do much of these catheterizations now days. So that's something we're thinking about how can we improve these patients? What are we missing?

Dr Carolyn Lam:                Yeah, if I could add my two cents. So Wouter mentioned finding the right therapies that can effectively reduce NT-proBNP safely, and well you mentioned choosing the right patients to use this in. And if I may, you know, just adding perhaps the right settings as well. Because it's well known that not all of us take care of heart failure patients the same way. And maybe there are settings where having a number to guide us may be more useful than others. But what do you do? You know, we wait for more data, but in the meantime, just congratulations. Heartfelt, heartfelt congratulations Wouter for a beautiful study, thank you so much for the privilege of publishing it in Circulation.

                                Thank you for being on this podcast, and listeners don't forget to tune in again next week.