LCZ696 is a heart failure drug from Novartis that many observers think will gain FDA approval later this year and go on to become a blockbuster. Perhaps the biggest obstacle to the drug’s success is the fear that it might raise the risk of Alzheimer’s disease. Now a new article in a top cardiology journal lays out the detailed basis for this concern. The authors do not contend that the Alzheimer’s issue will likely ruin the drug’s chances, but they do maintain that the problem needs to be carefully monitored.

I spoke with Milton Packer, the co-principal investigator of PARADIGM-HF, the mega trial that set off the mega excitement about the drug last year. He offers several persuasive arguments that Alzheimer’s disease won’t be the Achilles’ heel of LCZ696.

But first let’s look at the paper, published in the European Heart Journal, by a group of French researchers. The paper focuses on the potential impact of sacubitril, which is one of the two key ingredients of LCZ696. Sacubitril inhibits the action of neprilysin. The French authors focus on two conflicting roles for neprilysin. The cardiovascular benefits are achieved because in the heart neprilysin inhibition leads to increased levels of neuropeptides and vasodilators that are beneficial in hypertension and heart failure. However, in the brain neprilysin appears to play a key role in degrading the beta amyloid plaque that is one of the hallmarks of Alzheimer’s disease. The French authors raise the possibility that the drug could accelerate the progression of Alzheimer’s disease, particularly in people at high risk for the disease, though they confess that it is uncertain whether LCZ696 can penetrate the blood brain barrier, which would be necessary for this effect.

The French authors note that although no signal for cognitive dysfunction was found in PARADIGM-HF, the trial did not rigorously assess this endpoint. This endpoint will be followed much more closely with serial cognitive tests in the next big LCZ696 trial, PARAGON-HF, in heart failure patients who have preserved ejection fraction. However, write the authors, it is unclear “whether the duration of follow-up of this trial will be sufficient to exclude any deleterious effect of LCZ696 on cognitive function.”

The French authors acknowledge some important points that works against their concern. The second component of LCZ696, the angiotensin-receptor blocker valsartan, may protect against Alzheimer’s because of its vasoprotective properties. It is possible, therefore, that any negative effect of sacubitril could be counterbalanced by a positive effect of valsartan. It is also possible that the long time frame for Alzheimer’s development could mean that any adverse effect would be undetectable or clinically meaningless, given the advanced age of most heart failure patients.

Milton Packer Responds

The LCZ696 investigators are of course aware of these concerns. Animal studies, Packer told me, first raised the issue 10-20 years ago. “Fortunately,” he said, “we have actual data from long-term clinical trials (not only PARADIGM-HF but also OVERTURE), showing no increase (and actually a decrease) in dementia-related adverse events in patients receiving long-term treatment with neprilysin inhibitors.” (Remember that OVERTURE was the trial testing a similar agent, omapatrilat. It failed spectacularly, but not because of neurocognitive issues.)

Packer offered the same possible explanation as the French investigators for the observed decrease in dementia events. “The reason might be related to the fact the dementia is multifactorial and has an important vascular component, and the vascular component would actually be expected to be improved by neprilysin inhibition,” said Packer.

And here’s the important piece of news: “In any case,” Packer told me, “there was no imbalance (not even a numerical imbalance) in dementia-related events with omapatrilat or LCZ696.”

But there’s one important caveat: this data has not been presented or published yet. Packer told me the PARADIGM investigators have placed a “high priority” on the neurocognitive paper but it is important to reserve judgement until the data undergoes outside scrutiny. (And, not to be coy about it, Milton Packer has not always received a lot of praise for the timely publication of important papers.)

Packer also explained why PARAGON-HF will include a much more rigorous neurocognitive evaluation than PARADIGM-HF:

PARADIGM-HF focused on patients with a reduced ejection fraction, whose major morbidity and mortality is related to cardiovascular causes. The event rates are high; reducing them is really meaningful and trumps everything else. There was no reason to even consider dementia as a relevant issue.

PARAGON focuses on patients with a preserved ejection fraction, whose major morbidity and mortality is NOT related to cardiovascular causes. The event rates are lower; the patients are much older, and dementia is potentially a more important issue.

So it is important to have a sensitive instrument and analysis plan for dementia in patients with a preserved ejection fraction; not so in patients with a reduced ejection fraction.

In essence, if you have a drug that reduces mortality and hospitalizations for heart failure or for any reason convincingly, the decision to prescribe the drug does not (and will not) depend on the results of some dementia score or hypothetical concerns about dementia. Death trumps everything else. But without knowing the results of PARAGON, it is not possible to make the same assumptions.

This is not a time frame issue; it is a competing risk-to-benefit issue. Simply put, if you are dead or suffering from repeatedly hospitalizations for heart failure, you are probably not too worried about a small hypothetical effect on some dementia scale. No observational study will change the clear-cut decision for the patient.

I would just offer one thought here. Packer’s comments are entirely reasonable but he may be downplaying a potential future conundrum. The inclusion of a rigorous neurocognitive assessment in PARAGON raises the possibility of a troubling finding. (It’s much easier to find a signal for a problem that you are specifically looking for than if you are just generally trying to monitor adverse events, especially something that is extremely common in elderly patients.) I would imagine that a signal of neurocognitive harm in PARAGON might have a negative impact on the view of LCZ696 even in the very different population of PARADIGM patients. It would certainly lead to many awkward questions about the drug.

Outside Experts Weigh In

I asked several other experts to comment on the EHJ paper. Heart failure expert Clyde Yancy wrote that it is important that we “both temper our enthusiasm for new discoveries and mute our concerns when untoward effects might be a consideration.” We should “be resolute in our surveillance to insure safety first when new therapies are introduced.” But, he emphasized, “our enthusiasm for a new treatment paradigm in heart failure, and importantly the hope that so many patients have for improved outcomes should not be doused by what are at best  theoretical considerations.”

Clinical trial expert Sanjay Kaul agreed with Packer and the French authors that the potential adverse effects of sacubitril on amyloid beta might be counterbalanced by the drug’s beneficial vascular effects. He said he was somewhat reassured by the absence so far of an increase in dementia adverse events in OVERTURE and PARADIGM.