The Thirty-Six-Hour Gap

The prescribing information for sacubitril/valsartan contains an unusual instruction. When a patient switches between it and an angiotensin-converting enzyme inhibitor, at least 36 hours must separate the two treatments. Taking them together increases the risk of angioedema. The swelling can involve the face or lips and may obstruct the airway when it affects the tongue, glottis or larynx.^[1]

Thirty-six hours can look like a routine scheduling precaution. The reason for it reaches back to an earlier drug whose pharmacology proved too dangerous. Starting with the current label and working backward shows how several programs changed the idea that eventually became sacubitril/valsartan.

The case matters to Aeviant because it gives a rare public view of how a failed program can change the design that follows. The prescribing instructions for the final medicine include a precaution derived from that history.

Sacubitril/valsartan inhibits neprilysin. ACE inhibitors act on angiotensin-converting enzyme. Both enzymes help metabolize bradykinin and other vasoactive peptides. An overlap between the treatments temporarily inhibits both routes of clearance.^[1,2]

Omapatrilat once combined that pharmacology in a single experimental molecule. It inhibited ACE and neprilysin at the same time. The aim was to reduce signalling through the renin–angiotensin system while preserving endogenous natriuretic and vasodilatory peptides. The approach made sense in cardiovascular physiology, where those systems exert opposing effects.

The clearest warning came from OCTAVE, a trial that randomised 25,302 patients with hypertension to omapatrilat or enalapril. Omapatrilat lowered blood pressure more and helped more patients reach their treatment targets. It also produced angioedema in 2.17% of patients, compared with 0.68% receiving enalapril. Two omapatrilat-treated patients experienced airway compromise. Both recovered, although one required mechanical airway support.^[3]

Omapatrilat clearly had pharmacological activity. It reached its intended enzymes and lowered blood pressure. The safety problem came from the breadth of that action. Inhibiting both enzymes restricted overlapping routes for peptide clearance, including routes involved in bradykinin metabolism. The safety signal was linked to pharmacology central to the design.

The heart-failure results had already shown a different limitation. OVERTURE compared omapatrilat with enalapril in 5,770 patients with chronic heart failure. For the primary composite of death or heart-failure hospitalisation requiring intravenous treatment, the hazard ratio was 0.94. Omapatrilat met the prespecified criterion for non-inferiority, but not superiority. A prespecified secondary analysis and a post hoc analysis using a more conventional definition of heart-failure hospitalisation favoured omapatrilat, but neither changed the primary result.^[4]

Earlier still, a 48-patient, 12-week dose-ranging study had found that higher omapatrilat doses were associated with lower systolic blood pressure and dose-dependent changes in left-ventricular ejection fraction and end-systolic wall stress.^[5] The early pharmacology was sufficiently coherent to justify the larger program. OVERTURE later produced an insufficient efficacy result, and OCTAVE identified an unacceptable safety burden, even though the molecule remained pharmacologically active.

In early discovery, target engagement and movement in an expected physiological measure can justify further work. Omapatrilat achieved both. The larger program showed that its efficacy and safety could not be separated well enough for the treatment to be useful. Describing the molecule only as active or inactive would lose the information that shaped the next design.

The reason ACE and neprilysin inhibition were combined becomes clearer in the earlier work on neprilysin inhibition alone.

Neprilysin degrades several biologically active peptides, including natriuretic peptides released by the heart. Those peptides promote sodium excretion, vasodilation and other effects that oppose volume and pressure overload. In heart failure, reduced circulation activates systems that constrict blood vessels and retain fluid. Preserving natriuretic-peptide activity offered a way to strengthen a counter-regulatory response already present in the body.

Candoxatril, a selective neprilysin inhibitor, tested that idea clinically. In 110 patients with mild-to-moderate chronic heart failure who were already receiving ACE-inhibitor therapy, candoxatril increased mean total exercise time by 34.1 seconds relative to placebo over 84 days. Functional class, clinical status and quality-of-life scores did not differ significantly.^[6] The clinical signal was modest and appeared against a background of established treatment.

The physiology experiments showed why the response was difficult to predict. In two crossover cohorts of eight healthy volunteers each, candoxatril increased the plasma concentrations reached during exogenous angiotensin II infusion at both studied doses. The lower, non-natriuretic dose also enhanced the pressor response. The higher dose did not reproduce that effect.^[7] Neprilysin inhibition was changing the handling of signals with opposing cardiovascular consequences. Every relevant substrate of the enzyme remained part of the response.

Omapatrilat was developed in response to this problem. Neprilysin inhibition could preserve natriuretic peptides while increasing the influence of angiotensin II, so adding ACE inhibition offered a direct way to restrain the angiotensin system. ACE also participates in peptide clearance. Combining the two actions placed additional pressure on a clearance system that neprilysin inhibition had already changed.

Candoxatril and omapatrilat failed for different reasons. Candoxatril showed that selective neprilysin inhibition changed a wider substrate system than the original therapeutic summary suggested. Its clinical benefit was limited even when patients were already receiving ACE inhibition. Omapatrilat then showed that building ACE inhibition into the same molecule created a serious problem with overlapping bradykinin metabolism.

At Aeviant, failure is followed by a second question: what failed? The biology, the proposed mechanism, the molecule, the assay and the model are not interchangeable. A result may close the original program while revealing that the molecule is better suited to a different mechanism or therapeutic problem. That possibility is investigated rather than discarded. Predefined advancement and termination criteria still govern the original question. They prevent a program from being rescued by moving the goalposts after the result arrives. They do not require the information inside a failed result to be thrown away. The work can be redirected when the evidence gives it somewhere credible to go.

The next program kept neprilysin inhibition and used a different way to control the angiotensin system.

Sacubitril is converted in the body to sacubitrilat, or LBQ657, which inhibits neprilysin. It is paired with valsartan, an antagonist of the angiotensin II type 1 receptor. Angiotensin II can still be formed through ACE, while valsartan limits much of its signalling through the receptor responsible for many vasoconstrictive and sodium-retaining effects.^[1] This controls the angiotensin system without continuous ACE inhibition.

Receptor blockade left ACE available to participate in bradykinin degradation. The combination retained neprilysin inhibition and avoided the continuous ACE and neprilysin overlap built into omapatrilat. Angioedema remains a risk, and sacubitril/valsartan still carries warnings and contraindications for it. The revised design was sufficiently different to justify another clinical test.

Mechanism-led discovery cannot become molecule-loyal discovery. A molecular design is one attempt to answer a research question. Results from candoxatril and omapatrilat gave researchers a basis for retaining neprilysin inhibition and changing how the angiotensin system was controlled. The next design followed from the limitation each program had exposed.

An eight-week hypertension trial reported in 2010 provided an early clinical test. The study randomised 1,328 patients, of whom 497 received the combination then known as LCZ696. Compared with matched doses of valsartan, the combination produced greater blood-pressure reduction. No angioedema was reported among LCZ696 recipients during the study, although a sample of 497 patients observed for eight weeks could not exclude an uncommon risk.^[8]

The decisive evidence came from PARADIGM-HF. The trial randomised 8,442 adults with symptomatic chronic heart failure and reduced left-ventricular ejection fraction. Before randomisation, participants entered sequential run-in periods during which they received enalapril and then sacubitril/valsartan. Only patients who tolerated the required doses of both treatments proceeded to the double-blind comparison.^[1,9]

After a median follow-up of 27 months, cardiovascular death or first hospitalisation for heart failure had occurred in 21.8% of patients receiving sacubitril/valsartan and 26.5% receiving enalapril, corresponding to a hazard ratio of 0.80. Cardiovascular mortality, heart-failure hospitalisation and all-cause mortality were each lower in the sacubitril/valsartan group, and the trial was stopped early after crossing the prespecified boundary for overwhelming benefit.^[1,9]

The run-in remains important when interpreting that success. It selected a randomised population that had already tolerated target doses of both treatments, so adverse-reaction rates during the double-blind phase were lower than might be expected in unselected clinical practice. Hypotension occurred more frequently with sacubitril/valsartan than with enalapril. Angioedema remained a recognised risk, and kidney function, potassium and blood pressure still required clinical attention.^[1] The trial established an outcome benefit in a defined population. The treatment continued to carry important trade-offs.

PARADIGM-HF did not provide a complete molecular explanation for the benefit. Neprilysin processes numerous substrates, and their individual contributions are difficult to isolate inside a successful multicomponent treatment. In an open, uncontrolled mechanistic study of 73 patients switched to sacubitril/valsartan, soluble neprilysin activity fell while atrial natriuretic peptide increased approximately fourfold. Substance P and total amidated glucagon-like peptide 1 also increased. Plasma BNP and BNP activity did not change, while NT-proBNP declined modestly. The investigators proposed that atrial natriuretic peptide may have been a more important natriuretic-peptide effector than BNP in that cohort.^[10]

The clinical benefit is established more firmly than the contribution of any individual peptide.

Aeviant is still working at the stage where these distinctions have to be made prospectively. Its current research is computational, and no proposed mechanism has been confirmed experimentally. Structural models and molecular simulations help characterise targets. Docking can help prioritise hypotheses for testing. A coherent model can earn the right to be tested, but it cannot borrow the authority of an experiment it has not survived. Binding, functional activity and the resulting physiological state require direct experimental evidence. The computational work is intended to make those experiments more informative, including by defining results that would cause a program to change or stop.

Candoxatril showed that neprilysin inhibition remained worth studying, although its overall effects were difficult to predict from one substrate. Omapatrilat identified an unacceptable risk when ACE and neprilysin were inhibited together. Those results gave researchers a basis for pairing sacubitril with an angiotensin-receptor blocker. Sacubitril/valsartan later outperformed enalapril in heart failure.

The current Entresto label does not recount that development history. It gives the clinically relevant conclusion: allow at least 36 hours between an ACE inhibitor and sacubitril/valsartan. The requirement remains because simultaneous ACE and neprilysin inhibition still matters.