You are here

Improving heart failure outcomes: what is the role and future of new versus current treatments?

John McMurray (UK)

John McMurray examines pharmacological strategies to improve outcomes for heart failure patients, including questions health-care providers should ask before they replace an old treatment with a new one.



Improving heart failure outcomes: What is the role and future of new versus current treatments? - John McMurray

Improving heart failure outcomes: What is the role and future of new versus current treatments? - John McMurray

Thank you very much, Aldo. Good afternoon ladies and gentlemen.


These are my disclosures, you can see that I am a clinical trialist and work with lots of different treatments for heart failure.

What’s new in the 2016 guidelines?

Today the 2016 ESC Heart Failure Guidelines were published online, so were the American College of Cardiology American Heart Association Guidelines and there was an attempt to harmonise these guidelines so that the interpretation of the evidence and the recommendations on both sides of the Atlantic were as similar as possible.

Positive drug, device and other trials 2001-2016

In terms of drug therapy and, in fact, indeed in terms of drug and device therapy, there is only one single new treatment intervention described compared to the 2012 European, 2013 American guidelines, and that new intervention is based on the PARADIGM Heart Failure trial published in 2014, which came at the end of a sequence of really remarkable studies with drugs and devices conducted, this slide shows you approximately the past 15 years, but really it is a 30 year history of astonishing progress for our patients and Martin has mentioned many of those treatments.

Angiotensin receptor neprilysin inhibition (ARNI)

What was PARADIGM Heart Failure about? PARADIGM Heart Failure was a study looking at this new treatment, an angiotensin receptor neprilysin inhibitor. That is a compound consisting of two active moieties, one is a angiotensin receptor blocker, which everybody knows well, and the other is an neprilysin inhibitor, and not everybody is as familiar with neprilysin, the enzyme, and neprilysin inhibitors.

First-in-class” drugs improving outcome in heart failure

This trial made a lot of news when it was presented for the first time, because it was the first trial for about 15 years to show a new class of drug that reduces all-cause mortality in heart failure, the first treatment in about five years to show an improvement in morbidity and mortality, and that is quite a list of treatments that we now have and we have got to choose between these and how to add them together and so on, and that is what I am going to talk about.

When to replace an old treatment with a new one? Questions you should ask

Now I am very biased, so everything I say you have to listen to very critically, I was a Co-PI of the PARADIGM Heart Failure trial, I talked in many meetings about the PARADIGM Heart Failure trial, I think it was a very important trial, so I am coming at this from a biased perspective, but equally I tried to set out a series of questions that I think anybody should ask when they have to think about what this new guideline is recommending, which is to substitute sacubitril/valsartan for an ACE inhibitor in patients who remain symptomatic on good evidence-based treatment.

This is my list of questions, I am happy at the end if anybody wants to challenge these, but I think this is a reasonable list of questions and what I am going to do is to go through these and try and see how the PARADIGM Heart Failure trial and sacubitril/valsartan match up to the questions that are being asked.

When to replace an old treatment with a new one? Questions you should ask

The first thing – and this goes back to what Martin talked about – is was the benefit shown in PARADIGM Heart Failure clinically meaningful and in particular does it matter to patients?

PARADIGM-HF: Primary outcome – Cardiovascular death or heart failure hospitalization

This was the primary outcome; it was a composite of cardiovascular death or heart failure hospitalization. As you can see, this was reduced in relative terms by 20%, so sacubitril/valsartan was better than placebo. Now, does this outcome matter? Is it clinically meaningful?

PARADIGM-HF: The primary outcome

This is a hospital in Glasgow, this is Glasgow Royal Infirmary, a very old hospital, built beside the cathedral, beside where the university was, it is a medieval hospital and like very old, medieval hospitals and cathedrals it is situated right beside a graveyard. This is a cemetery, patients in the wards here can think about what might happen in the future, so I would argue that the primary endpoint was highly clinically relevant, hospital admission and death. This is sadly the reality for many patients with heart failure and these are clinically meaningful endpoints as well for us, as doctors, for society, for healthcare systems, because those hospital admissions are expensive.

PARADIGM-HF: Components of the primary outcome

When you look at a composite endpoint you should also critically look to see did the components of that composite respond in the same way to the treatment? A composite is made up of different bits and it is very important that the different components show the same directional effect, at least, of treatment, and you can see here that both cardiovascular death and heart failure hospitalization were reduced approximately equally.

PARADIGM-HF: Cause/mode of death

You might say, if I am the patient, ‘I don’t care whether I die from a cardiovascular cause or whether I die from a non-cardiovascular cause, it is dying that is important.’ In this trial we also saw a reduction in all-cause mortality, there is a 16% relative risk reduction in all-cause mortality.

Now some of your more sophisticated patients may actually be aware of the modes of death that they are likely to experience. I can remember one time when we had just got the evidence about implantable cardioverter defibrillators reducing the risk of death and I spoke to one of my patients who was a retired anaesthetist anaesthesiologist and I told him about this device, and I told him he would be eligible for that and he said to me ‘No, thank you, I have seen patients with heart failure during my career, I have seen what happens when people die from progressive pump failure, if I die suddenly I will be quite happy about that.’ Even patients who know a lot about heart failure may have their own thoughts about mode of death and you can see here that we reduced both of those two key modes of death in heart failure, sudden death and death from worsening heart failure.

PARADIGM-HF: Number of hospital admissions (including repeats) by cause

You may also say ‘Well, actually patients with heart failure don’t just get admitted to hospital with worsening of their condition, they get admitted with all sorts of things and sometimes it is actually very difficult to tell what is actually the cause of the admission.’ I see that all the time, how many patients do you see it seems to be worsening heart failure and maybe there is a chest infection, maybe it was their atrial fibrillation, so we know that and in fact we know that the majority of hospital admissions in patients with heart failure are not actually due to worsening of the heart failure, but, again as you can see in this slide, in our trial we saw a reduction in all types of admissions, we saw a reduction in heart failure, for sure, we saw a reduction in cardiovascular admissions and we even saw a reduction in all causes of admissions.

PARADIGM-HF: First and repeat heart failure hospitalizations

When you look and when I show you and when people stand up on podiums like this and present results of trials, they show you time to first event, those Kaplan-Meiers are the time to the first one of those events that a patient experiences, but you and I know that our patients with heart failure are frequently readmitted, often multiple times, and again you can see here in the PARADIGM Heart Failure trial that that is exactly what we saw, even though these patients had, as Martin said, by and large mild symptoms. Over the 27 months median follow-up time in PARADIGM Heart Failure they were often admitted two or even three or sometimes four times, but, again as you can see from this slide, we not only reduced first admissions, but we did prevent recurrent admissions in these patients.

PARADIGM-HF: Intensive care management

One of the most intriguing things – and I still don’t, at least in my own mind, clearly understand this – is that even if patients were admitted to hospital during the trial you can see from this slide that they seemed to be less sick, so we saw fewer patients admitted in the trial in the sacubitril/valsartan group who required intensive care treatment, we saw fewer days spent in intensive care, we saw less use of intravenous inotropic agents, so even the patients who were admitted, if they were in the sacubitril/valsartan group, seemed to have less severe heart failure on admission.

When to replace an old treatment with a new one? Questions you should ask

I would say those endpoints really do matter to patients, but Martin said, and I think he is right, that for some patients their quality of life is extremely important as well and I would argue that being admitted to hospital, at least in Glasgow, isn’t necessarily the best quality of life experience that you can have, but, of course, the majority of a patient’s life lived with heart failure is lived in the community, it is not lived in hospital.

Kansas City Cardiomyopathy Questionnaire (KCCQ)

We did measure quality of life in PARADIGM Heart Failure in exactly the same way as was measured in the SHIFT trial, as mentioned by Martin, and that was using the Kansas City Cardiomyopathy Questionnaire, a questionnaire that has 23 different items, 23 questions, covering five domains of a patient’s life. It asks about their symptoms, it asks about what they can do, it asks about their quality of life, it asks about their ability to live a normal life in interacting with their friends, meeting with their family members, being able to travel and so on. If you have done trials you may have used this and this may be familiar to you.

KCCQ: Significance of a 5-point change

Now I am going to show you a post hoc analysis, so this was not the pre-specified primary analysis, but what I am going to do is show you what you might call a responder analysis, so this is the proportion of patients having a five or greater point decrease in their KCCQ score.

Now why am I looking at a five point change? The reason I am doing that is that has been defined many years ago as a clinically important change, so a five or more point reduction in score equates to a clinically significant deterioration and, indeed, in some studies this has been correlated with, for example, changes in 6-minute walk distance, changes in peak VO2 on the treadmill, so this is widely regarded as being clinically important.

PARADIGM-HF: Percentage of patients with at least 5 points deterioration in KCCQ scores at month 8

This slide shows you in the same way as all the other slides the enalapril group in black, the sacubitril/valsartan group in the hatched blue bars and this is all of these different domains of a patient experienced existence with heart failure, looking at their symptoms, looking at their physical limitation, their social interaction, whatever you want to look at, and if you look across this slide you will see in every single domain there was a statistically significantly lower risk of showing deterioration if the patient was randomised to sacubitril/valsartan than randomised to enalapril.

PARADIGM-HF: Percentage of patients with at least 5 points deterioration in KCCQ clinical summary score

Here is not just month eight, which was the pre-specified time that we would look at this, in the trial, but month 12 and month 24 and you can see, again, there was a persistent benefit in quality of life and, if you like – and I will come back to these – numbers needed to treat, then you would need to treat about 23 patients to prevent one having a clinically important deterioration in their quality of life by eight months.

I believe, but you can challenge me, that those endpoints are clinically important to patients. As Martin said, we should try and ensure it in our trials.

When to replace an old treatment with a new one? Questions you should ask

My second question about any trial is, were the results robust? Were they statistically convincing?

PARADIGM-HF: Statistical robustness

Here I do feel on fairly firm ground, because I don’t know if you noticed the P value for the primary endpoint, but it was 4 x 10-7, that is equivalent to having four or five individual trials significant at the P less than 0.05 level. If you would like to put this another way, it equates to there being a less than 1 million chance that these findings could have been due to the play of chance.

When to replace an old treatment with a new one? Questions you should ask

I believe the endpoints were clinically important, I believe the results were statistically robust, but was the benefit really worthwhile?

PARADIGM-HF: Primary outcome – Cardiovascular death or heart failure hospitalization

This is always interesting, because this is how we look at results, usually we look at Kaplan-Meier curves, we look at relative risk reductions, hazard ratio, so a hazard ratio of 0.8, a relative risk reduction of 20%, but we are more and more encouraged to look at absolute risk reductions and there are various ways to do that.

PARADIGM-HF: Primary outcome – Cardiovascular death or heart failure hospitalization

One is a number needed to treat, so what this means is that during the 27 months of median follow-up in PARADIGM Heart Failure I would have to switch 21 patients from an ACE inhibitor to sacubitril/valsartan to stop one of them having cardiovascular death or heart failure hospitalization.

PARADIGM-HF: Absolute benefits

Another way to look at this is the number of events per 100 or per 1,000 patients treated. This slide shows you the absolute benefits that we would have obtained by switching 1,000 patients from an ACE inhibitor to sacubitril/valsartan and, again, you can see that these are pretty large absolute treatment benefits and I will come back to that, because, of course, that plays into cost effectiveness and, again, we must always ensure that we get value for money with the treatments that we use.

PARADIGM-HF: estimated long-term event-free survival gain at various ages

Then another way to think about drugs, particularly drugs that will be given for a few years, as they will these days in patients with heart failure, is how long can you extend life?

Brian Claggett, who is a Statistician working with Scott Solomon in Boston, has done some really nice work trajecting extension of life in patients with heart failure in the PARADIGM trial, and here you see, if you were to look at patients aged 65 years of age when they started treatment, if they had been switched from an ACE inhibitor to sacubitril/valsartan you would see, for example, that the extension of life is 1.3 years, extension of event-free survival 1.6 years – it is almost identical, by the way for a 70 year old individual.

PARADIGM-HF: estimated long-term event-free survival gain at various ages

How does that compare to some of the things that we know and love in cardiology?

Many years ago Salim Yusuf did one of the very first analyses of this sort, a really beautiful analysis that opened my eyes to thinking about what we do when we use treatments, and he looked at the extension of life obtained with coronary artery bypass surgery, because by that stage the coronary artery bypass trials had long-term follow-up. Coronary artery bypass surgery for patients with three vessel disease extends life by about half a year, coronary artery bypass surgery for left main coronary disease extends life for about 1.6 years; very similar to what I have just shown you for sacubitril/valsartan.

When to replace an old treatment with a new one? Questions you should ask

The next question that I have put up here reluctantly, because it probably isn’t a question I generally ask, but it is one that I am frequently asked by others and that is, is the benefit with the new treatment seen consistently across subgroups in the trial?

PARADIGM-HF: Sub-group analysis (primary endpoint and CV death)

The answer was ‘Yes’. I am hoping you can’t read any of this and that is deliberate, but I will show you a simpler subgroup analysis and this is for me to explain to you why I am uncomfortable about talking about subgroups.

PARADIGM-HF: Subgroups

Does anybody have any comment – this is for audience interaction now – does anybody think that any of these subgroups look different? Do you think this subgroup here, number 7, looks as though perhaps they don’t do as well on sacubitril/valsartan as on enalapril? That would be what many people would say, if you asked them this question.

PARADIGM-HF: Subgroups: Month of birth

These are patients in PARADIGM sub-grouped according to their month of birth and I am only showing you this, hopefully, to make you believe that subgroups can, by the play of chance, lead to findings that are spurious. I am just hoping nobody here thinks that they can explain why sacubitril/valsartan doesn’t work in people born in July.

The elderly

But I think one group of people that we should look at are the elderly and I think we should look at the elderly because they reflect the patients we see, at least in Western Europe and in North America, in practice, but often patients in trials are younger. Now people point this out all the time and so they should, but I would just point out to you that in a trial like PARADIGM, conducted in 47 countries worldwide, the lower average age to a great extent reflects the fact that patients in Latin America and in the Asia Pacific region are much younger than patients in Western Europe and North America. The average age of patients in Western Europe and North American PARADIGM was between 68 and 72 years of age, but the elderly, as Martin showed, are neglected. He showed very, very clearly how the elderly don’t get given treatments. I don’t believe it is because of intolerance, because there are plenty of data to show that the elderly, although they might not tolerate the full dose of treatment, actually tolerate most of the drugs that we want to use very well, it has probably got as much to do with the doctors and the nurses looking after the patients as it has to do with the patients themselves.

Efficacy and tolerability of sacubitril/valsartan according to age

We have looked at this issue in PARADIGM Heart Failure and one of my colleagues in Glasgow did this very nice analysis, I think it is a great way to look at something like age, a continuous variable.

Benefit of sacubitril/valsartan according to age in PARADIGM-HF

Here are two outcomes, heart failure hospitalization, all-cause mortality, and this blue line represents a continuous hazard ratio, so this is the sacubitril/valsartan and enalapril hazard ratio, you will remember this was 0.8 overall, so 0.8 down here, this is the line of unity, so that would indicate no difference between the treatments, and these shaded areas represent 95% confidence intervals, so down here, at a very young age, these are extremely wide, because we had hardly any patients in their twenties in PARADIGM Heart Failure. As you can see, the 95% confidence intervals get a bit wider up here, because we had very few very elderly patients, but the reason this stops here is because the oldest patient in PARADIGM Heart Failure was 96, and if you look at this continuous hazard ratio for heart failure hospitalization it remains below 1 right across the spectrum of age.

If you look at all-cause mortality, okay, maybe in your late eighties/early nineties we are not going to reduce your mortality very much, but right across most of that spectrum of age you can see the hazard ratio is below unity.

PARADIGM-HF: Effect of LCZ696 according to age category

This tells you that there was a consistent benefit across the spectrum of age and not only was there a reduction in the hospitalization and mortality, but there was also a smaller proportion of patients in the older age group who experienced a clinically important reduction in quality of life. I think that is crucially important because perhaps it is in the oldest people where quality of life is at least as important, if not more important, than duration of life, and in those oldest patients we saw an improvement in quality of life.

When to replace an old treatment with a new one? Questions you should ask

Is the benefit of the new treatment truly incremental? I think that is an important question to ask. Martin pointed out to you the many, many different effective treatments we have and, of course, you want a new treatment to be tested on top of those existing treatments. There is no point in showing a new treatment does something good if patients are not getting the current, existing best treatment.

PARADIGM-HF: Incremental benefits of sacubitril/valsartan: CV death

Here you see the hazard ratios for patients with sacubitril/valsartan compared with enalapril according to a number of different background therapies, diuretics, digoxin, mineralocorticoid receptor antagonist, defibrillating device, even if the patient had previous coronary artery bypass grafting, and no matter what treatment subgroup we looked at we saw a benefit of sacubitril/valsartan over enalapril.

Additional benefit of sacubitril/valsartan in patients taking an MRA: Cardiovascular death

But the one that was most important for me was the mineralocorticoid receptor antagonist subgroup, because only about half of the patients, just over half of the patients, in PARADIGM were on an MRA and that was because we didn’t know patients with mild symptoms should be on an MRA when we started the trial, but, as you can see here, these patients, who remember were all getting a RAS-blocker, were almost all getting a beta-blocker, 93%, and even if in addition they were getting an MRA, they still had a benefit from switching from an ACE inhibitor to sacubitril/valsartan.

Incremental benefits of sacubitril/valsartan according to background β-blocker use

Both Aldo and Martin mentioned dose of drug. We know that doctors and nurses, for some reason, under dose patients with heart failure. Again, if we looked at beta-blocker dose then whether patients were on a bigger dose or a lower dose they still benefit from sacubitril/valsartan.

When to replace an old treatment with a new one? Questions you should ask

That is all about benefits, but, of course, the patient must also be able to take the treatment, the treatment must be well tolerated, it must not have important adverse side effects.

PARADIGM-HF – Prospective comparison of ARNI with ACEI to Determine Impact on Global Mortality and morbidity in Heart Failure trial

These are a little bit more difficult to understand in PARADIGM Heart Failure, because we had an active run-in period, so everybody had to tolerate a reasonable dose of ACE inhibitor, then tolerate sacubitril/valsartan, before they were randomised.

PARADIGM-HF: Pre-specified safety endpoints

But after randomisation here you can see we saw more hypotension, we actually saw less renal impairment, we saw less hyperkalaemia, we were worried about angioedema, but we did not see a significant amount of angioedema.

PARADIGM-HF: Adverse events leading to permanent study drug discontinuation

I think the most remarkable thing was that after randomisation actually fewer patients in the sacubitril/valsartan group stopped treatment for an adverse event than in the enalapril group and, most importantly, the one adverse effect that was definitely more common with sacubitril/valsartan than with enalapril, that was hypotension, that did not lead to significantly more withdrawals in the sacubitril/valsartan group than in the enalapril group and, in fact, very few patients had to stop [the] study drug because of hypotension.

I am nearly finished.

When to replace an old treatment with a new one? Questions you should ask

What about ease of use, drug interactions, monitoring and so on? This is a twice daily drug, it doesn’t have any significant drug interactions that we know of, nor any significant food interactions and the monitoring is really the same as you would have to do for an ACE inhibitor, ARB, beta-blocker or mineralocorticoid receptor antagonist.

When to replace an old treatment with a new one? Questions you should ask

To finish with, the last question is, is the treatment cost effective?

Sacubitril/valsartan: Cost-effectiveness

I am not a health economist, I have already said I am biased, so instead of giving you my opinion I am going to turn to what we have in the United Kingdom and Martin talked about England, the United Kingdom is more than England, there is this little bit at the top called Scotland, and in Scotland we have something called the Scottish Medicines Consortium, that is like our version of NICE in England. Both of these organisations act as the payor for the National Health Service and they do a cost effectiveness analysis and unless your drug is cost effective, unless it meets a certain cost per a quality adjusted life year threshold then it will not be approved. I sit on one of these committees, we are forever seeing cancer drugs that cost enormous amounts of money, they get turned down by NICE and the next day the Daily Mail or some other English newspaper has a big story about how the Government rations healthcare in the UK, but NICE and the Scottish Medicines Consortium both approved sacubitril/valsartan as a cost effective treatment that should be funded within the UK National Health Services.

There was another independent report in the United States that also, I believe, considered sacubitril/valsartan to be a cost effective treatment.

When to replace an old treatment with a new one? Questions you should ask

Those are the questions I posed. I am very happy to be challenged about whether you think those were fair questions to ask, I have given you data, at least, to address the answers.

Treatment of HF-REF: the current situation ........

That was yesterday, before the new guidelines were published.

Treatment of HF-REF: The future?

Maybe this will be today, tomorrow and the future.

Thank you very much indeed. [Applause]

This site is sponsored by:

Search this site


Subscribe to our E-Alert to keep up to date with the new items in the Resource Centre.