Cardiologists have been happily prescribing angiotensin receptor neprilysin inhibitor, sacubitril-valsartan, for heart failure with reduced ejection (HFrEF) for a while now. And the drug was recently approved by the FDA for heart failure with ejection fraction, “below normal”, taking into account the results of PARADIGM-HF in HFrEF, and also reviewing the evidence from PARAGON-HF in HF with preserved ejection fraction.
Novartis also wanted to capture the post-MI market. Their double-blind randomized trial, PARADISE-MI, of twice daily sacubitril-valsartan started within 7 days versus ramipril for this indication is due to report soon.
Which post-MI patients did PARADISE-MI recruit?
It recruited thousands of people with spontaneous myocardial infraction. Their ejection fraction had to be 40% or less and they had to have pulmonary congestion requiring intravenous treatment associated with the index MI. On top of those criteria (i.e. already quite a sick patient), they had to have at least one of the following 8 risk factors:
- Age ≥ 70 years
- eGFR <60 mL/min/1.73 m^2 based on MDRD formula at screening visit
- Type I or II diabetes mellitus
- Documented history of prior MI
- Atrial fibrillation as noted by ECG, associated with index MI
- LVEF <30% associated with index MI
- Worst Killip class III or IV associated with index MI requiring intravenous treatment
- STEMI without reperfusion therapy within the first 24 hours after presentation
But at the same time they had to be haemodynamically stable in the 24 hours prior to randomization which was defined as a systolic blood pressure greater than 100mmHg and no intravenous diuretics, vasodilators, or inotropes. Suffice it to say, the majority of MIs would not meet these strict criteria. However, there is a rationale.
Firstly, in practice, it is challenging to start sacubitril-valsartan (or indeed ramipril) if the blood pressure is lower than that and one has to wait. Secondly, by selecting very high risk patients, the investigators increase the chance of detecting lots of events in the follow-up period. This increases the power of the study to detect a benefit of the drug if there is a one.
What is outpatient heart failure?
The primary endpoint was time to one of the following:
- Cardiovascular death
- Heart failure hospitalization
- Outpatient heart failure
Secondary endpoints (which will be assessed hierarchically) include subsets of the primary endpoint components, as well as all-cause mortality, stroke and MI. A long list of exploratory endpoints is also provided. So far so good.
I haven’t come across “outpatient heart failure” in a primary endpoint before. The reason we see hospitalization for heart failure in endpoints is because needing to be admitted signifies severity, and because it relatively easy to determine whether it happened. But it doesn’t sound unreasonable, to want to know if a drug reduces heart failure regardless of whether it requires admission. Heart failure that would tip a frail person into hospital means just as much as the same pathology in someone less frail. And by including milder events in the endpoint, the power of the study is increased.
- An urgent/unscheduled visit to an ED, acute/urgent care facility or outpatient clinic or a non-urgent office/practice or study visit for a primary diagnosis of HF that does not require an overnight hospital stay.
- Patients must exhibit at least one documented new HF symptom with objective evidence of clinical HF consisting of at least 2 physical examination findings or one physical examination findings and at least one laboratory criterion.
- The event requires initiation or intensification of treatment specifically for HF. Such treatment can include administration of intravenous agent (e.g., diuretic, vasodilator, vasopressor, or inotrope) or mechanical or circulatory intervention for HF, OR initiation of oral loop diuretic treatment, or intensification of oral maintenance loop diuretics for the diagnosis of HF, over a sustained period (i.e., initiation or doubling of total daily dose through a period of ≥ 4 weeks), which is confirmed at a subsequent outpatient visit
Novartis’ finance report from the first quarter of 2021 states: “While numerical trends consistently favored Entresto in a head to head comparison with ramipril, a current standard of care, the Ph3 PARADISE-MI study did not meet its primary composite endpoint of reducing risk of cardiovascular death and heart failure events after an acute myocardial infarction. The safety profile of Entresto was confirmed. Novartis will continue to evaluate the data. Topline results will be presented at the American College of Cardiology 70th Annual Scientific Session”.
So, one may ask why the benefits seen in heart failure weren’t replicated here? I am afraid we will not know until the full results are published.
That report is not intended to report the full trial results, but to comply with Securities and Exchange Commission by announcing the topline result, which is “the Ph3 PARADISE-MI study did not meet its primary composite endpoint of reducing risk of cardiovascular death and heart failure events after an acute myocardial infarction” . Anything else is speculative at this stage, and we will need to wait until the full results of the trial are published. Journals enable results to be fully explained and put into context in the main trial publication, and actually, likely in many more secondary publications. Therefore, the press release either said too little or it said too much, stating that “numerical trends consistently favored Entresto in a head to head comparison with ramipril”. It’s legitimate to perform subgroup analyses to understand how the drugs works and which patients could benefit. Is there an interaction between renal function and benefit of the drug perhaps? Or between type of MI (STEMI vs NSTEMI) and the benefit of the drug? In fact, they’ve pre-specified 23 subgroups to look at. So I’d be interested to see what these analyses show, along with the main findings of course! It is not unusual also to look into components of the composite primary endpoint, and also into secondary and exploratory endpoints, which represent much weaker evidence.
Faiez Zannad comments: “In order to understand why the benefits seen in heart failure, up to a normal ejection fraction weren’t replicated in post MI patients with low ejection fraction and/or signs and symptoms of heart failure, we will need to read Table 1: i.e patients characteristics. The devil is in the details. The main inclusion criterion was indeed “Evidence of LV systolic dysfunction (EF<40%) and/or pulmonary congestion requiring intravenous treatment”. How many patients had low EF and how many had simple (transient?) congestion signs? We must wait until we know the full story!
Interestingly, ACE inhibitors will remain unchallenged in this indication of post-MI, low EF and/or signs of congestion. VALIANT showed that an angiotensin receptor blocker is not superior to captopril, and OPTIMAAL showed that losartan was not superior to captopril. Now, the preliminary understanding of PARADISE MI is that sacubitril/valsartan is not superior to ramipril. The interesting finding here is that neprilysin inhibition (sacubitril) does not seem to add much, to the effect of valsartan. Only the mineralocorticoid receptor antagonist eplerenone, in EPHESUS, added to the benefit of ACEi and ARBs.
In order to have an understanding of the totality of evidence we need to wait until we see the totality of the data.”