Pitfalls of Therapies Targeting the Nitric Oxide Signaling Pathway in Heart Failure With Preserved Ejection Fraction

Abstract:

Over 30 million individuals globally are afflicted with heart failure with preserved ejection fraction (HFpEF). Despite this substantial figure, treatment options for HFpEF remain limited. This is largely attributed to the variability in disease features and comorbidities with which patients present in the clinic. A common feature among patients with HFpEF is reduced nitric oxide (NO) bioavailability—a finding which prompted the use of therapeutics targeting the NO signaling pathway in HFpEF preclinical and clinical trials. Although many of these therapeutics were successful in animal models, clinical trials have yielded neutral, negative, or contradictory results. In this review, we will summarize the outcomes of HFpEF clinical trials investigating drugs that target the NO signaling pathway (nitrates, nitrites, soluble guanylyl cyclase stimulators, and phosphodiesterase 5 inhibitors) and discuss potential pitfalls underlying the neutral or negative results, as well as considerations for the design of future studies.

INTRODUCTION

Heart failure with preserved ejection fraction (HFpEF; see Table 1 for all abbreviations and acronyms) affects over 30 million individuals worldwide, accounting for half of all heart failure cases.¹ This condition, often accompanied by symptoms of exercise intolerance and dyspnea, is defined by a left ventricular ejection fraction of ≥50% and evidence of 1 or more structural and/or functional changes in the heart that suggest the presence of diastolic dysfunction (eg, elevated natriuretic peptides, increased left atrial volume, increased left ventricular mass index, elevated E/e’, and/or increased pulmonary artery systolic pressure).² HFpEF can be more specifically categorized as primary or secondary based on the respective absence or presence of other cardiac complications (eg, valvular, myocardial, and/or pericardial diseases) that can lead to increased left ventricular filling pressures.³ This distinction between primary and secondary HFpEF is important to make, as both the prognosis and course of treatment for patients vary largely between these 2 disease classifications. With respect to prognosis, it has been shown that age- and sex-matched primary and secondary patients with HFpEF have comparable rates of all-cause mortality.⁴ However, unsupervised clustering of patients with HFpEF (primary and secondary combined) based on age, sex, comorbidity status, left ventricular ejection fraction, and natriuretic peptide levels revealed all-cause mortality is reduced in the resulting cluster of patients containing a majority (∼69%) of primary patients with HFpEF relative to the other patient clusters that were enriched in secondary patients with HFpEF (3-year cumulative all-cause death of ∼35% vs. ∼50%).⁴

TABLE 1.

Relevant acronyms and abbreviations used throughout the text,

Acronyms and Abbreviations
CAPACITY-HFpEF	A study of the effect of IW-1973 on the exercise capacity of patients with heart failure and preserved ejection fraction
cGMP	Cyclic guanosine monophosphate
Cpc-PH	Combined pre- and post-capillary pulmonary hypertension
DILATE-1	Acute hemodynamic effects of riociguat in patients with pulmonary hypertension associated with diastolic heart failure
haemoDYNAMIC	Riociguat in pulmonary hypertension and heart failure with preserved ejection fraction—haemoDYNAMIC trial
HFpEF	Heart failure with preserved ejection fraction
INABLE-Training	Inorganic nitrite to amplify the benefits and tolerability of exercise training in heart failure with preserved ejection fraction
INDIE-HFpEF	Inorganic nitrite delivery to improve exercise capacity in HFpEF
Ipc-PH	Isolated post-capillary pulmonary hypertension
KCCQ PLS	Kansas City Cardiomyopathy Questionnaire Physical Limitation Score
KNO3	Potassium nitrate
KNO3CKOUT HFpEF	Effect of KNO3 compared with KCl on oxygen uptake in heart failure with preserved ejection fraction
MPMA	Matching perfusion and metabolic activity in HfpEF
NaNO2	Sodium nitrite
NEAT-HFpEF	Nitrate's effect on activity tolerance in heart failure with preserved ejection fraction
NO	Nitric oxide
NT-proBNP	N-terminal pro-B-type natriuretic peptide
PASSION	Phosphodiesterase-5 inhibition in patients with heart failure with preserved ejection fraction and combined post- and pre-capillary pulmonary hypertension
PDE5	Phosphodiesterase-5
PDE5i	Phosphodiesterase-5 inhibitor
PH-HFpEF	Pulmonary hypertension secondary to heart failure with preserved ejection fraction
RELAX	Phosphodiesterase-5 inhibition to improve clinical status and exercise capacity in diastolic heart failure
RONS	Reactive oxygen–nitrogen species
RV	Right ventricle
SAK HFpEF	SGLT2i and KNO3 in HfpEF
sGC	Soluble guanylyl cyclase
SOCRATES-PRESERVED	Soluble guanylate cyclase stimulator in patients with heart failure and PRESERVED ejection fraction
TOPCAT	Treatment of preserved cardiac function heart failure with an aldosterone antagonist
VITALITY-HFpEF	Evaluate the efficacy and safety of the oral sGC stimulator vericiguat to improve physical functioning in daily living activities of patients with heart failure and preserved ejection fraction
VO2	Volume of oxygen consumption

In regard to treatment, secondary HFpEF can be treated by targeting the underlying cardiac disease leading to diastolic impairment.³ Conversely, primary HFpEF—where aging and cardiometabolic comorbidities (eg, hypertension, obesity, and type II diabetes) contribute to intrinsic diastolic dysfunction of the left ventricle—should be treated using SGLT2 inhibitors,^5,6 or, as of more recently, GLP-1 agonists,⁷ which have been shown to improve patient clinical status. Although both SGLT2 inhibitors and GLP-1 agonists have shown great therapeutic promise in primary patients with HFpEF, these therapies have only just begun to be recognized for their use in HFpEF over the last 5 years. The prior deficit in effective therapeutics occurred, in part, due to the heterogeneity of HFpEF and an incomplete understanding of its pathophysiology. The current theory surrounding primary HFpEF etiology is that a reduction in nitric oxide (NO) bioavailability (arising from inflammation and oxidative stress derived from comorbidities) underlies the structural and functional changes occurring across multiple organs.⁸ This, in turn, has driven a clinical focus on targeting the NO signaling pathway in patients with HFpEF over the last decade.

Targeting the NO signaling pathway has been shown to be efficacious in animal models of HFpEF.^9–13 However, such studies have not translated similarly in HFpEF clinical trials. In this review, we will summarize the efficacy of drugs targeting each node of the NO signaling pathway (Fig. 1) in patients with HFpEF, highlighting potential pitfalls of the existing studies and describing future considerations for targeting the NO signaling pathway in HFpEF.

FIGURE 1.

Drug targets in the nitric oxide signaling pathway. By the conventional pathway, endothelial nitric oxide synthase (eNOS) produces nitric oxide—which may also be pharmacologically supplemented by nitrate or nitrite—in endothelial cells from substrates, l-arginine and oxygen. Nitric oxide diffuses to neighboring vascular smooth muscle or cardiac myocytes and binds its receptor, soluble guanylyl cyclase (sGC), to produce cyclic guanosine monophosphate (cGMP). cGMP activates protein kinase G (PKG), promoting signaling involved in smooth muscle relaxation and cardiac compliance and suppressing hypertrophic signaling in cardiac myocytes. To boost cGMP levels, sGC stimulators can be used as well as inhibitors of phosphodiesterase 5 (PDE5), which negatively regulates nitric oxide signaling by degrading cGMP. Figure produced using BioRender.com.

NITRIC OXIDE PRODUCTION, SIGNALING, AND IMPACT ON THE CARDIOVASCULAR SYSTEM

Nitric oxide was first identified around 40 years ago as the endothelium-derived relaxing factor released in response to endothelial-dependent vasodilators.^14,15 Since then, NO production and action in cardiovascular health and disease have been studied and subsequently reviewed in great detail.^16,17 Nitric oxide can be produced through 2 pathways. In the first, more traditional pathway, NO synthases utilize oxygen and l-arginine as substrates to produce NO.¹⁸ In an alternative pathway, known as the nitrate–nitrite–NO pathway, NO is produced from the reduction of its metabolites, nitrate and nitrite, in hypoxic and acidotic settings.^19,20 Upon its production, NO acts through 1 of 2 mechanisms. In the first, NO binds soluble guanylyl cyclase (sGC), producing cyclic guanosine monophosphate (cGMP) to activate protein kinase G and downstream signaling (Fig. 1).²¹ In the other, NO directly modifies proteins at sites containing sulfhydryl groups to not only form stabler, bioactive S-nitrosothiols but also modulate protein activity.^22,23 Through these mechanisms, NO acts on cells of the cardiovascular system (eg, endothelial cells, vascular smooth muscle cells, and cardiac myocytes) to regulate both vascular and cardiac homeostasis. Among the many functions of NO signaling in cardiovascular health, those most relevant to the subject of this review include regulating vascular function (promoting endothelial health and driving smooth muscle relaxation), depressing hypertrophic signaling in the heart,²⁴ and maintaining cardiac compliance.^25–27 As NO plays such central roles in the cardiovascular system, it is important that its bioavailability be maintained.

NITRIC OXIDE SIGNALING IN HFpEF

The concept that NO bioavailability is reduced during HFpEF is well accepted and has been shown in both patients with HFpEF and animal models of the disease.^{10,26,28–30} This phenomenon is considered to arise from a series of events at the onset of HFpEF progression. The presence of systemic inflammation during HFpEF leads to increased generation of reactive oxygen–nitrogen species (RONS), namely superoxide.⁸ Superoxide readily combines with NO—lowering NO bioavailability—to form the harmful RONS, peroxynitrite, as evidenced by elevated tyrosine nitration in human HFpEF myocardium.^26,31 As with other RONS, homeostatic/baseline levels of peroxynitrite can reversibly modify proteins to facilitate beneficial redox signaling.³² For example, in the presence of glutathione, peroxynitrite modifies cysteine residues (through S-glutathiolation) of sarco/endoplasmic reticulum calcium (Ca²⁺) ATPase (commonly known as SERCA) to increase calcium re-uptake and smooth muscle relaxation.³³ Once homeostatic levels are surpassed (as occurs in conditions associated with oxidative stress), chronic overproduction of peroxynitrite results in irreversible nitration or oxidation of proteins (as well as lipids and nucleic acids), which can deleteriously alter protein activity and promote cellular stress.^34,35 An example of this in HFpEF is peroxynitrite-mediated oxidation of the endothelial NO synthase cofactor, tetrahydrobiopterin, which leads to uncoupling of the synthase.^30,34–36 Once endothelial NO synthase becomes uncoupled, the enzyme's capacity to produce NO is diminished, and instead the synthase favors production of superoxide, which further contributes to the oxidative, NO-depleted environment observed in HFpEF.^31,36

Consequences of decreased NO bioavailability during HFpEF are extensive and likely contribute to the onset of symptoms, such as exercise intolerance, in patients. At the vascular level, blood vessels are less responsive to stress (ie, exercise), blunting their ability to adequately perfuse and deliver oxygen to tissues.³⁷ This can be particularly problematic in the vasculature of skeletal muscle, which has a high oxygen demand during exercise. Vascular dysfunction can also increase systemic blood pressure and subsequent cardiac afterload, leading to pathological cardiac remodeling in the form of left ventricular hypertrophy. Decreased NO bioavailability can also drive structural and functional changes in the heart through direct action on the heart. Nitric oxide signaling has been shown to prevent the development of cardiac hypertrophy, with NO deficiency driving the enlargement of cardiac myocytes.²⁴ Furthermore, supplementation with protein kinase G is sufficient to reduce the passive stiffness in cardiac myocytes derived from patients with HFpEF, suggesting a deficiency in NO signaling directly impairs cardiac compliance.²⁶ Together, these consequences of low NO bioavailability that occur during HFpEF point to the NO signaling pathway as a promising therapeutic avenue.

THERAPIES TARGETING NITRIC OXIDE SIGNALING

Nitrate Therapy

Nitrates have been used therapeutically for cardiovascular disease for hundreds of years. As early as the eighth century, Buddhists reported using what is now known to be inorganic nitrates to treat angina.³⁸ Centuries later, the use of inorganic nitrates to treat cardiovascular disease led to the development of organic nitrates, such as nitroglycerin, isosorbide mononitrate, and isosorbide dinitrate, which are currently prescribed to treat cardiovascular diseases, including angina, myocardial infarction, and heart failure with reduced ejection fraction. Whether nitrate therapy (inorganic or organic) is effective in the treatment of HFpEF has only begun to be studied within the last decade. We will describe the results from organic and inorganic nitrate trials separately, as the two classifications of nitrate therapy vary in mechanism and therapeutic impact.

Organic Nitrates

Organic nitrates are commonly prescribed vasodilators in many cardiovascular diseases, especially those that require quick relief (eg, angina). This is due to the fast-acting and highly potent nature of these synthetic compounds. Upon delivery to the bloodstream, organic nitrates are converted to NO within minutes by enzymes, such as cytochrome P450 or mitochondrial aldehyde dehydrogenase, and are cleared from the circulation after 30 minutes.³⁹ Although lower dosing of this therapy is sufficient to drive peripheral venodilation and reduce cardiac preload, higher doses, used in more severe conditions (eg, heart failure, angina, myocardial infarction), drive arterial dilation to reduce cardiac afterload and increase coronary vascular dilation and perfusion.⁴⁰ As such impacts have the potential to offer clinical benefit during HFpEF, multiple trials have been conducted over the last decade to determine whether organic nitrates are effective in treating HFpEF.

Unexpectedly, among all trials examining the efficacy of organic nitrates in HFpEF, it has been shown that patient clinical status is not improved but rather worsened with nitrate use. In the earliest study, named the NEAT-HFpEF trial (NCT02053493), patients were administered isosorbide mononitrate (once daily at 30 mg, 60 mg, then 120 mg) for 6 weeks.⁴¹ Patients taking nitrates had reduced mean arterial pressure and systolic blood pressure, though these reductions did not lead to any improvement in exercise capacity. Rather, isosorbide mononitrate worsened exercise capacity, leading to a trending (P = 0.06) reduction in daily activity levels, as well as decreased hours of daily activity—results that were unexpected given the beneficial effect of organic nitrates on blood pressure. A second study conducted 2 years later by Zamani et al⁴² showed somewhat similar results (NCT01516346). In this trial, patients were administered isosorbide dinitrate (40 mg thrice daily) with or without an inhibitor of superoxide biosynthesis, hydralazine (75 mg thrice daily), for 6 months. Those receiving isosorbide dinitrate alone had lower systolic blood pressure and increased arterial compliance relative to the placebo group, with no change in exercise capacity as measured by 6-minute walking distance.

Both of the described trials showed that organic nitrates were not only ineffective in improving exercise capacity in patients with HFpEF but also poorly tolerated, as drug-related adverse events were increased in those taking nitrates.^41,42 Such events or side effects, such as hypotension and lightheadedness, accounted for half of the subject dropouts in the latter study.⁴²

A study by Chirinos et al⁴³ helped to explain this poor tolerability. They showed nitroglycerin (0.4 mg; administered once) had preferential effects on the cerebral vasculature—carotid vessels had increased cross-sectional diameter and reduced resistance in response to nitroglycerin. Meanwhile, no change in wave reflections arriving at the aorta was measured. The authors attribute the localized effect of nitroglycerin to neurons being enriched with mitochondria and, therefore, rich in aldehyde dehydrogenases to drive the metabolism of the drug. This could, in turn, help explain the side effects (eg, lightheadedness) experienced by patients and potentially the neutral effects elsewhere.

Two other studies examined long-term outcomes in patients with HFpEF taking organic nitrates. In the first, Lim et al⁴⁴ used patient data from the SwedeHF database to showed patients with HFpEF taking nitrates had an increased risk of all-cause mortality or first heart failure hospitalization (hazard ratio: 1.35). However, this effect was lost when comparing across a matched cohort. A second study by Tsujimoto et al later examined patient data from the TOPCAT trial, which tested the efficacy of aldosterone antagonism in HFpEF.⁴⁵ From data collected during that trial, the authors observed that nitrate use increased the risk for both major cardiovascular adverse events (hazard ratio: 1.32) and heart failure hospitalization (hazard ratio: 1.32). Together, these currently published studies indicate that organic nitrates are not just ineffective in treating HFpEF but also detrimental.

Pitfalls of Organic Nitrate Trials

There are multiple explanations as to why the described organic nitrate trials may have resulted in neutral or deleterious outcomes in patients with HFpEF. The mechanism by which organic nitrates are predicted to work is through the supplementation of NO to the vasculature, which could, in turn, reduce blood pressure and cardiac afterload. Although the studies that measured blood pressure at rest observed reductions in patients taking nitrates, these studies did not measure vascular function during exercise.^41–43 Impairment in vasodilatory reserve is a key feature associated with HFpEF and a major contributor to exercise intolerance.⁴⁶ Thus, it is possible that nitrate use was sufficient to reduce blood pressure at rest but not during exercise, accounting for the lack of improvement in exercise capacity observed in these patients (though not necessarily explaining the adverse effects observed in some studies). The impact of organic nitrates on vascular function during exercise is therefore worth considering if further investigations ensue in the context of HFpEF.

Another important consideration that may explain the results of the organic nitrate trials is the tolerance to therapy that can arise from frequent and/or chronic nitrate use. Nitrate tolerance is a well-established concept specific to organic nitrates in which patients have a reduced response to therapy when taken chronically. Although the mechanism of nitrate tolerance is not completely understood, it has been shown that RONS produced from chronic nitrate use can increase NO scavenging and inhibit the function of both aldehyde dehydrogenase and sGC to prevent nitrate bioactivation and downstream NO signaling.⁴⁷ In line with this idea, inhibiting superoxide biosynthesis with hydralazine prevented nitrate tolerance in otherwise healthy rabbits.⁴⁸ Although this mechanism helps to explain the few neutral observations made in organic nitrate HFpEF trials, there is another mechanism of nitrate tolerance that may account for the negative outcomes measured across most HFpEF studies. Chronic nitrate use has been shown to induce vascular expression of endothelin-1,⁴⁹ a vasoconstrictor that has been reported to already be elevated in the serum of patients with HFpEF independent of therapeutic intervention.⁵⁰ An increase in endothelin-1 biosynthesis not only could directly drive vessel constriction but could also increase vascular sensitivity to other vasoconstrictors, such as norepinephrine and angiotensin II.^47,49 Neurohormonal activation is a well-established feature in HFpEF.⁵¹ Thus, organic nitrates may prime the vasculature to be hyperresponsive to the pathological neurohormonal environment and subsequently less able to dilate in response to stressors, like exercise. This idea would help to explain the results of the study by Zamani et al,⁴² which showed a negative impact of hydralazine in patients with HFpEF taking organic nitrates, which the authors attributed to potential sympathetic activation from hydralazine administration. Nitrate tolerance can be circumvented with the use of an extensive daily nitrate-free period (∼10 hours a day).³⁹ However, the NEAT-HFpEF trial used a once-daily dose for 6 weeks, and a negative impact was still observed with nitrate use.⁴¹ Thus, although the timing of doses is important in preventing nitrate tolerance, it may not be sufficient when organic nitrates are administered over a long duration and may not explain the unfavorable outcomes observed.

Thus far, studies examining the status of patients with HFpEF taking organic nitrates have only been investigated following chronic use. It remains unclear whether less frequent use of organic nitrates, specifically only on exercise-heavy days, would allow for an acute benefit of organic nitrates without the development of tolerance. Treatment of HFpEF is likely to require chronic therapy, so it is crucial that timing and dosing be optimized in the future to prevent tolerance if there is to be any potential for success in organic nitrate therapy for HFpEF.

Inorganic Nitrates

Despite similarly acting to produce NO, inorganic nitrates vary largely from organic nitrates in how they are processed and what impacts they produce during HFpEF. Inorganic nitrates, which are exogenously derived from one's diet (ie, leafy greens) or through pharmacological preparations, are converted to NO through the nitrate–nitrite–NO pathway, which has been described in great detail.^39,52 In brief, ingested inorganic nitrate is absorbed in the upper gastrointestinal tract and delivered to the bloodstream, with plasma nitrate levels elevating as early as 30 minutes and peaking within 3 hours.³⁹ Approximately 25% of the circulating nitrate becomes concentrated in the salivary glands, where it can be excreted into the oral cavity and reduced to nitrite by commensal anaerobic bacteria harboring nitrate reductases.⁵² It is important to note that human cells lack a nitrate-reducing enzyme, so the action of inorganic nitrates is reliant on an intact oral microbiome.⁵² Following its production, nitrite is absorbed through the upper gastrointestinal tract into the bloodstream, where it reaches peak levels within 2.5 hours.³⁹ Nitrite can then be converted to NO through various enzymes, including globins (hemoglobin, hemoglobin alpha,⁵³ myoglobin), xanthine oxidoreductase, NO synthases, and certain mitochondrial proteins.^39,52 NO is preferentially produced from nitrite in hypoxic and acidotic contexts⁵² (eg, in skeletal muscle during exercise), offering a controlled-release system targeted to durations of exercise when patients with HFpEF are most symptomatic.

Over the last decade, multiple clinical trials have ensued to determine the efficacy of inorganic nitrate therapy in patients with HFpEF. In the first trial (NCT01919177), Zamani et al⁵⁴ showed that a single acute dose of beetroot juice (12.9 mmol nitrate in 140 mL) reduced vascular resistance during exercise and improved exercise capacity (peak oxygen uptake [VO₂], total work performed, and time to exhaustion) in patients with HFpEF. Aortic pressure-flow data from this study were published 2 years later, showing that inorganic nitrates reduced the arterial wave reflection coefficient and delayed the onset of wave reflections.⁴³ These results suggest that fewer wave reflections arrived at the proximal aorta and that those that did likely reached the proximal aorta when the heart was in diastole. These changes would reduce the late systolic pulsatile load the heart experiences and, in the long term, ameliorate the structural and functional changes in the heart that may have originally occurred due to increased afterload. However, this study was too short to measure such alterations in the heart.

In between the aforementioned 2 studies, a second study was published that also displayed a beneficial effect of inorganic nitrate use in HFpEF. This study by Eggebeen et al⁵⁵ tested the efficacy of a lower dose of nitrate (6.1 mmol nitrate in 70 mL beetroot juice) taken as either a single acute dose or a once-daily therapeutic for the duration of 1 week. Although both durations yielded a decreased systolic blood pressure at rest, only the week-long regimen was efficacious at reducing systolic blood pressure at exhaustion (Δ = −7 mm Hg, P = 0.054) and increasing time to exhaustion during exercise. Individuals from this study were also enrolled to determine whether inorganic nitrates and exercise training would have a synergistic effect on exercise capacity.⁵⁶ Relative to individuals who were exercise trained without nitrate therapy, those exercising and taking nitrates displayed no additional improvement in functional status. However, plasma nitrite levels were not significantly increased in those on nitrate therapy, which may account for why no additional benefit was observed.

Although earlier studies used beetroot juice to administer inorganic nitrates, multiple trials utilized pharmacological preparations of inorganic nitrates (ie, potassium nitrate; KNO₃). Zamani et al⁵⁷ conducted the first study examining both the safety and efficacy of KNO₃ (6 mmol, twice daily for 1 week, then thrice daily for 1 week) in patients with HFpEF (NCT02256345). Unlike organic nitrates, KNO₃ was well tolerated. KNO₃ also reduced systolic blood pressure at rest but had no effect on vasodilatory reserve or peak VO₂. Other metrics of exercise capacity were improved (eg, exercise duration, total work performed), though these end points were exploratory. Successive blood draws were taken from these patients for pharmacokinetic studies, which may account for the lack of improvement in exercise capacity. Most recently, a larger study called the KNO₃CKOUT HFpEF trial (NCT02840799) examined the efficacy of chronic KNO₃ use (6 mmol, thrice daily for 6 weeks) in HFpEF.⁵⁸ This study reported a reduction in mean arterial pressure during exercise without any change in vasodilatory reserve or exercise capacity.

Pitfalls of Inorganic Nitrate Trials

The inconclusive results of the KNO₃CKOUT HFpEF trial were somewhat surprising given the acute benefit of inorganic nitrates in HFpEF. As stated by the authors, it is possible that compensatory mechanisms arose from chronic use of inorganic nitrates, which negated the acute benefit of the therapy.⁵⁸ However, there are other variables that may explain this outcome, as well as the outcomes of the other neutral studies. A major point of concern is whether nitrates were properly metabolized into nitrite in the trials in which no effect was observed. Although NO metabolites were elevated in all studies, not all studies reported individual measures of circulating nitrite. This leaves the possibility that nitrite levels were unchanged in those studies, which would help to explain the neutral results observed. When nitrite levels were reported to be elevated, an improvement in both blood pressure during exercise and exercise capacity was achieved.⁵⁵ However, in multiple studies, nitrite levels did not significantly change, and subsequently, no change in blood pressure during exercise nor exercise capacity was observed.^56,57 Given the hypothesized mechanism by which inorganic nitrate is proposed to improve exercise capacity is through the conversion of nitrite to NO during exercise, the lack of elevated nitrite levels is highly problematic. The KNO₃CKOUT HFpEF trial did not report individual nitrate and nitrite levels but instead generally indicated that NO metabolites (a single readout representing combined levels of nitrate, nitrite, NO–metal and heme complexes, thiol–NO adducts, and protein cysteine–NO adducts) were elevated.⁵⁸ As they used a similar dosing scheme of KNO₃ that was used in a previous study that yielded no elevation in nitrite levels,⁵⁷ it is possible that the neutral results of the KNO₃CKOUT HFpEF study can be attributed to improper metabolism of the study drug. As oral bacteria are the sole reducers of nitrate in humans,⁵² an altered oral microbiome in patients with HFpEF could account for the unchanged nitrite levels observed in some studies. It has been shown that patients with HFpEF have an altered gut microbiome relative to healthy controls.⁵⁹ Hyperglycemia or impaired glycemic control, which is considered a key component of metabolic syndrome and a contributor to HFpEF development, has also been linked to altered oral microflora.⁶⁰ However, whether there is variation in oral microbial populations among patients with HFpEF and whether there are functional deficits in nitrate reduction to nitrite have yet to be assessed.

Heart structure and function were mostly unchanged throughout these trials, which was an interesting finding, as the current paradigm for HFpEF is that reduced NO bioavailability drives hypertrophy and stiffening of the left ventricle.⁸ It is possible that the length of these studies was too short to detect a change in heart structure or function. Another possibility is that the heart itself does not meet the threshold of hypoxia and acidosis to drive a sufficient amount of local production of NO from nitrite.

Other factors that may account for the variability in the outcomes of these trials include the patient population and dosing. Most of these studies were conducted in small populations that varied in sex, race, and disease severity. The first successful trial occurred in a group composed largely of black men,⁵⁴ while all other studies (from which only 1 was successful⁵⁵) were predominantly made up of white women. Sex differences in the response to inorganic nitrates have been reported, with healthy men having an improvement in exercise capacity while healthy women experience little to no effect.⁶¹ However, postmenopausal women experience an improvement in vascular function when taking inorganic nitrates and exercising, suggesting that the absence of estrogen may confer a therapeutic benefit from nitrate therapy.⁶² Dosing of nitrates may further account for the lack of improvement in some studies. For example, 12.9 mmol nitrates was sufficient at improving exercise capacity with a single dose,⁵⁴ while 6.1 mmol was not unless taken daily for a week.⁵⁵ In the study by Shaltout et al,⁵⁶ no additional improvement in exercise capacity occurred in patients taking a 6.1 mmol dose administered thrice weekly with exercise. This lack of an effect could potentially be attributed to not taking nitrates as frequently as once daily.

Together, there are multiple explanations for the neutral outcomes in many of the clinical trials described in this section. Future efforts with this therapy should be aimed at testing and boosting nitrate conversion to nitrite, whether it be through dosing, timing, the nitrate source (beetroot juice vs. KNO₃), or oral microflora reestablishment. Thus far, studies using pharmacological preparations of KNO₃ have yet to specifically show increased nitrite levels in patients with HFpEF, which must be considered in the development of future trials.

Nitrite Therapy

As described in the previous section, inorganic nitrates may offer limited success in patients with HFpEF due to impairment in the patients' capacities to reduce nitrates into nitrites. To circumvent this issue, inorganic nitrites can be directly administered to patients with HFpEF. Multiple trials have been conducted over the last decade to determine the efficacy of doing so. In the first, Borlaug et al⁶³ showed that intravenous administration of sodium nitrite (NaNO₂; 250 μg/kg) acutely reduced both left (post-capillary wedge pressure) and right-sided (right atrial pressure) pressures in the heart, as well as pulmonary artery pressure in patients with HFpEF during rest and exercise (NCT01932606). Further, cardiac output reserve and VO₂ were improved, suggesting nitrite can provide an acute benefit in patients with HFpEF.

Because intravenous delivery of NaNO₂ is not feasible for chronic use, future studies tested whether aerosolizing dissolved NaNO₂ and delivering it through an inhalation device was similarly efficacious in patients with HFpEF. As with intravenous administration, nitrite administered using a nebulizer at a final dose of 90 mg acutely reduced post-capillary wedge pressure, right atrial pressure, and pulmonary artery pressure during rest^64–66 and exercise⁶⁵ in patients with HFpEF (NCT01431313, NCT02262078). Nebulized nitrite also increased pulmonary artery compliance,^64–66 potentially accounting for the improvement in right ventricular (RV) efficiency that was also observed.⁶⁶ A study combining data from previous trials described^63,65 revealed that receiving nitrite intravenously versus through an inhalation device achieved comparable outcomes relative to their respective controls.⁶⁷ This study, combining intravenous and nebulized data, showed nitrite increased cardiac output and peak VO₂ in patients with HFpEF. Improvements in oxygen transport in the peripheral and pulmonary circulation were also achieved by nitrite. The successes in these initial trials led to the examination of chronic administration of nebulized NaNO₂ (46 mg thrice daily for a week, then 80 mg thrice daily for 3 weeks) in patients with HFpEF.⁶⁸ This study, known as the INDIE-HFpEF trial (NCT02742129), revealed NaNO₂ conferred minimal clinical benefit for patients–blood pressure (mean arterial and diastolic) was reduced, though no differences in exercise capacity or heart failure status (ie, heart filling pressures, N-terminal pro-B-type natriuretic peptide [NT-proBNP] levels, left atrial index, or quality of life) were detected.⁶⁸ A recent study, named the INABLE-Training trial (NCT02713126), found a similar result in which administration of oral NaNO₂ (40 mg thrice daily for 12 weeks) with exercise training conferred no additional benefit in patients with HFpEF compared with those receiving exercise training alone.⁶⁹ These results preliminarily suggest that inorganic nitrites offer only a transient benefit for patients with HFpEF.

Pitfalls of Nitrite Trials

Both trials measuring the efficacy of chronic nitrite therapy in patients with HFpEF resulted in neutral outcomes.^68,69 A major concern acknowledged by the authors in the INDIE-HFpEF trial was drug delivery.⁶⁸ This group previously showed acute success with the delivery of aerosolized NaNO₂ in patients with HFpEF.⁶⁵ However, the nebulizer device used in their initial study was single-use, so they had to switch to a different device for the chronic trial. The new device not only may have differed intrinsically in its drug-delivering capacity, but had also been difficult for patients to use,⁶⁹ which could have further impeded drug administration. In line with the idea that drug delivery may have been impaired, plasma cGMP levels were unchanged between the treatment and placebo groups.⁶⁸ These measurements were taken at trough levels, so it is possible that levels were different directly following drug administration. Nitrite treatment was able to reduce blood pressure, which suggests that its levels were sufficient to drive some changes, though those levels may not have been sufficient to alter any other outcomes.

Another factor that may have led to the neutral outcome in the INDIE-HFpEF trial is the patient population recruited. As HFpEF is heterogeneous in how it presents, it is possible that certain subpopulations respond better to nitrite therapy. In 2 of the 4 acute studies showing success, patients were specifically recruited if they had pulmonary hypertension secondary to HFpEF (PH-HFpEF),^64,66 while patients were recruited for the INDIE-HFpEF trial with broader inclusion criteria.⁶⁸ Thus, it remains possible that chronic nitrite therapy may still be beneficial in patients with PH-HFpEF, especially given that the inhaled route of administration directly targets the pulmonary circulation. Future studies testing the efficacy of chronic nitrite delivery through an inhalation device in patients with PH-HFpEF should therefore be considered. However, the feasibility of such a study is uncertain, as the nitrite formulation used in these studies for nebulization has been discontinued.⁶⁹

Timing in the INDIE-HFpEF trial may have also accounted for its neutral outcome. In this study, patients took nitrite thrice daily with a minimum of 4 hours between each dose.⁶⁸ Although it has been determined that the half-life of nitrite administered through an inhalation device is approximately 30–40 minutes,⁶⁵ how long the effect of nitrite persists beyond that is unclear. One study in healthy subjects showed that the impact of oral Na¹⁵NO₂⁻ (20 mg) on blood pressure lasts for a maximum of 2 hours.⁷⁰ Whether this timing remains true when aerosolized nitrite is administered is unclear. If it does, the timing of doses in the INDIE-HFpEF trial would have continuously caused a return to baseline before the next dose was to be administered. Aside from the potential problem with the timing of doses, the INDIE-HFpEF trial was also a relatively short trial, lasting only 4 weeks. This timespan may have simply been too short to observe any lasting impact on exercise capacity and heart failure status.

The INABLE-Training trial showed there was no added benefit of nitrite use with exercise training in patients with HFpEF.⁶⁹ This, in part, may be explained by the dose of nitrite administered. This trial was interestingly the only HFpEF trial to administer nitrite orally, largely due to the nitrite formulation used for nebulization being discontinued early on into the study. The oral dose taken daily was 40 mg, which is much lower than the aerosolized dose (90 mg) given in the successful acute studies. Given that the oral route of administration makes drugs more susceptible to metabolism before reaching the bloodstream, it is possible that even less than the intended dose of nitrite was delivered to the circulation. However, no measurements of plasma nitrite were described in these patients, so it is unclear how much of the therapy was circulating or how that compared with nitrite levels in the prior successful studies. A higher dose of oral sodium nitrite (80 mg twice daily) than the dose administered in the INABLE-Training trial (40 mg thrice daily) resulted in a plasma nitrite concentration of 7 μM in otherwise healthy middle-aged and older subjects.⁷¹ This concentration is much lower than that measured after nitrite was administered through a nebulizer and therapeutic benefit was observed in patients with HFpEF (plasma nitrite concentration = 11.1 μM).⁶⁵ Therefore, it is possible that the dose of nitrite administered in the INABLE-Training trial was too low for a synergistic effect of nitrite and exercise training to be measured. More work is thereby warranted to determine the dose of oral nitrite required to achieve circulating drug levels that are comparable with those of the trials that were successful in patients with HFpEF.

The same dosing scheme and duration applied in the INABLE-Training trial were also used in a previous study, which examined the impact of oral nitrite on patients with metabolic syndrome and hypertension—two common comorbidities in HFpEF.⁷² Although this trial showed that oral nitrite was sufficient at reducing blood pressure (systolic, diastolic, and mean arterial pressure), the effect was lost after 8–10 weeks of nitrite therapy and returned to baseline by the 12th week.⁷² Together, these data point to the idea of nitrite tolerance, though the mechanism for this is unclear. Given that the INABLE-Training trial used the same dosing timeline, it is possible that a benefit derived from nitrite was missed, given that the end outcomes were measured after 12 weeks. Future work should carefully consider the study duration and also the potential use of a nitrite washout period after 8–10 weeks to determine whether that could recover the blood pressure response.

A final factor that may be preventing the success of chronic nitrite therapy in patients with HFpEF is the untreated cardiometabolic comorbidities that lead to reduced NO bioavailability in HFpEF. During these comorbidities, inflammation and oxidative stress occur, with RONS (ie, superoxide) from the latter potently scavenging NO and producing peroxynitrite. Thus, without the treatment of oxidative stress, the supplemented NO from nitrite may be scavenged before it can exert its action. Future work should therefore examine how nitrite therapy impacts patients with HFpEF when combined with an inhibitor of superoxide production, such as hydralazine, which was beneficial in murine HFpEF.¹⁰ Direct treatment of cardiometabolic comorbidities may also reduce both inflammation and oxidative stress to make the environment more amenable to NO supplementation. Thus, the pairing of nitrite therapy with a GLP-1 agonist or SGLT2 inhibitor (both of which target cardiometabolic comorbidities on top of treating HFpEF) could pose as a potential solution if RONS stemming from these diseases is, in fact, neutralizing the effect of nitrite.

Soluble Guanylyl Cyclase Stimulators

Several therapeutic strategies have been explored to elevate cGMP levels, either directly or indirectly, using organic and inorganic nitrates, nitrites, or phosphodiesterase inhibitors. However, these approaches have shown limited efficacy, often due to nitrate tolerance or insufficient endogenous cGMP production.^41,47,73 In this regard, the use of sGC stimulators, namely vericiguat, riociguat, and praliciguat, has been investigated in clinical trials as potential therapeutic agents for HFpEF and other cardiovascular diseases.⁷⁴ To date, there have been 3 double-blinded, placebo-controlled, phase II clinical trials in which patients with HFpEF have been treated with sGC stimulators, including (SOCRATES)-PRESERVED (NCT01951638),⁷⁵ VITALITY (NCT03547583),⁷⁶ and CAPACITY (NCT03254485),⁷⁷ and 2 studies on patients with PH- HFpEF (DILATE-1 (NCT01172756)⁷⁸ and haemoDYNAMIC (NCT02744339)).⁷⁹

DILATE-1 was the first study to use an sGC stimulator in patients with HFpEF. This trial included a 30-day study period following a single dose of riociguat to explore safety, tolerability, and pharmacokinetics. Unsurprisingly, with 39 patients split between a placebo group and 3 separate single-dose treatments, DILATE-1 did not meet its primary end point of reduced mean pulmonary artery pressure or pulmonary arterial wedge pressure. However, the study demonstrated that riociguat was well tolerated, increased cardiac output, and reduced diastolic blood pressure.⁷⁸ Soon after, the (SOCRATES)-PRESERVED trial, which evaluated vericiguat in 477 patients with HFpEF, showed mixed results.⁷⁵ Patients were randomized to receive placebo, vericiguat at fixed doses (1.25 mg or 2.5 mg daily) or vericiguat titrated to target doses (5 mg or 10 mg daily, starting from 2.5 mg) for 12 weeks with 2 co-primary end points: change in log-transformed NT-proBNP and left atrial volume. Neither of these primary end points was reached. Nonetheless, enthusiasm was sparked due to a significant finding of improved quality of life by the Kansas City Cardiomyopathy Questionnaire Physical Limitation Score (KCCQ PLS, +19.3 vs. +9.2 with placebo, P = 0.016), which included self-reported improvements in activities involving moderate physical exertion.⁸⁰ It is unclear whether these patient-reported outcomes fully translated to functional changes, as functional assessments, such as 6-minute walking distance or the grip strength test, were not measured during the study. The findings of improved quality of life measured in the (SOCRATES)-PRESERVED trial were not replicated in 2 subsequent clinical trials: VITALITY-HFpEF, which involved 789 patients across 21 countries,⁷⁶ and CAPACITY-HFpEF,⁷⁷ which consisted of 196 patients with HFpEF across 59 sites in the United States and Canada. The VITALITY-HFpEF trial included patients with chronic and recently worsening HFpEF who were randomized to receive vericiguat (either 10 or 15 mg daily) or placebo for 24 weeks. The primary outcome of improved KCCQ PLS was unmet as there was no significant improvement with either vericiguat dose compared to placebo. There was similarly no difference in the secondary end point, 6-minute walking distance. However, what did emerge from this study was a mortality imbalance, with more cardiovascular deaths occurring in vericiguat groups (8 cardiovascular deaths in the 15 mg/day dose and 12 cardiovascular deaths in the 10 mg/day dose compared with 4 cardiovascular deaths in the placebo). Despite this concerning increase in mortality measured in the vericiguat groups, small numbers preclude definitive conclusions on the overall safety of this therapy. The CAPACITY-HFpEF study, in which patients with HFpEF either received praliciguat (40 mg once daily) or placebo for 12 weeks, similarly failed to reach its primary end point of increased VO₂ max or its secondary end points, including improved 6-minute walking distance, ventilatory efficiency, and biomarkers such as NT-proBNP and troponin T. Moreover, the praliciguat group had a lower KCCQ score than the placebo group and a higher incidence of treatment-emergent adverse events. Meta-analyses of (SOCRATES)-PRESERVED, CAPACITY-HFpEF, and VITALITY-HFpEF have since further supported the lack of an effect of sGC stimulators on HFpEF outcomes, though a large-scale randomized study would be required to truly assess changes in all-cause mortality with sGC stimulators.⁸¹

Although the first 4 HFpEF-focused trials (DILATE-1, [SOCRATES]-PRESERVED, VITALITY-HFpEF, and CAPACITY-HFpEF) all failed to reach their primary end points, the recent haemoDYNAMIC study reached its stated primary end point of increased cardiac output, which was also seen in DILATE-1. However, there was no effect on more clinically relevant parameters, including pulmonary arterial wedge pressure, systemic vascular resistance, NT-proBNP levels, 6-minute walking distance, quality of life, or mortality.⁷⁹

Pitfalls of Soluble Guanylyl Cyclase Stimulator Trials

The trials examining sGC stimulators in HFpEF yielded mixed results, in part due to trial-specific limitations. For example, in the (SOCRATES)-PRESERVED trial, the primary end points of NT-proBNP levels and left atrial volume were unmet.⁷⁵ However, the 12-week study duration was likely too short to observe left atrial remodeling, especially with the dose escalation timeline, which only administered the end point doses in the titrated arms in the final 8 weeks.

(SOCRATES)-PRESERVED was also hampered by a dosing error, which excluded 48 patients across the 2 titrated dose arms (20 from the 5 mg group; 28 from the 10 mg group), lowering the sample size. Despite these limitations, the (SOCRATES)-PRESERVED trial observed improvements in quality of life based on KCCQ PLS, which was not recapitulated in later trials. VITALITY-HFpEF, which had an unmet primary end point of KCCQ PLS, used an a priori approach of assigning worst-case scores (score of 0 out of 100) to deceased patients.⁷⁶ The increased cardiovascular mortality observed in subjects randomized to receive vericiguat may have artificially deflated the apparent treatment effect. This study population also had a higher baseline KCCQ PLS and better functional status compared with the previous (SOCRATES)-PRESERVED group, potentially creating a ceiling effect that made symptomatic improvements difficult to observe, especially with a surprising 7.3-point increase in the KCCQ PLS in the placebo group.⁷⁶ Along a similar idea, there were concerns in the CAPACITY-HFpEF trial about patients' ability to perform a low-symptom cardiopulmonary exercise test to obtain peak VO₂ measurements, which likely resulted in the majority inclusion of patients with milder HFpEF.⁷⁷ These limitations, individual to each trial, may have prevented the measurement of positive outcomes in patients with HFpEF taking sGC stimulators.

On a molecular level, it is possible that sGC stimulators are ineffective at treating HFpEF based on their mechanism of action. This category of therapeutics stimulates sGC independently of NO by binding to the domain which NO typically binds heme within sGC—an interaction that is reliant on sGC containing ferrous (Fe²⁺) heme.⁸² However, during HFpEF and other cardiovascular diseases alike, oxidative stress is common and can lead to the oxidation of sGC to produce ferric (Fe³⁺) heme or complete loss of heme from sGC.⁸² Such alterations in sGC status would reduce, if not completely abolish, the action of sGC stimulators, which may explain the neutral outcomes in the described studies. Targeting oxidative stress (more specifically superoxide and/or peroxynitrite as the primary oxidants driving HFpEF) with antioxidants and/or treatment of cardiometabolic comorbidities may therefore offer a larger therapeutic effect with sGC stimulation and should be investigated in future work. Another avenue to circumvent the barrier of oxidative stress would be by using sGC activators (eg, ataciguat, cinaciguat), which can activate sGC in its oxidized or heme-deficient form.⁸² Currently, no work has been published on the use of sGC activators in HFpEF, revealing an area in need of further investigation.

It also must be noted that by therapeutically targeting sGC, vascular smooth muscle cells and cardiac myocytes—where cGMP exerts its action—would be most greatly impacted. This leaves endothelial cells essentially untreated, as sGC stimulators do not replenish NO which endothelial cells require for proper function. Endothelial dysfunction is thought to not only precede many of the structural and functional changes that occur in the heart during HFpEF but also dictate patient outcomes.^8,83 Therefore, it is possible that if endothelial cells are not targeted in HFpEF treatment, as is the case with sGC stimulators, it may remain difficult to measure greater therapeutic impacts.

As it stands now, the success of sGC stimulators in heart failure with reduced ejection fraction trials⁸⁴ has not been recapitulated in HFpEF studies, nor have the HFpEF trials yielded any consistent improvements in quality of life to suggest their use as a therapeutic for patients with HFpEF. Future work in this area should consider targeting oxidative stress alongside sGC stimulation or the use of sGC activators in the treatment of HFpEF.

Phosphodiesterase 5 Inhibitors

Phosphodiesterase 5 (PDE5) is an enzyme that negatively regulates NO signaling by way of degrading cGMP. Thus, as a mechanism to replenish NO signaling, PDE5 inhibitors (PDE5i), such as sildenafil and tadalafil, have been investigated in HFpEF clinical trials. The first study examining PDE5 inhibition in HFpEF, published by Guazzi et al (NCT01156636), was a small single-center trial, which revealed sildenafil (50 mg, thrice daily for 12 months) was successful at both reducing pressures and improving function of the pulmonary vasculature, right heart, and left heart. Consequently, patient quality of life also improved.⁸⁵ However, despite this early success, the Guazzi study was followed by multiple studies showing sildenafil conferred no benefit in patients with HFpEF.^73,86,87 Among these was the RELAX trial (NCT00763867)—a large, multicenter study in which patients were administered sildenafil (20 mg thrice daily for 12 weeks, then 60 mg thrice daily for 12 weeks) for a total of 24 weeks.⁷³ Not only did this study result in no alteration in cardiovascular function or exercise capacity with PDE5i therapy, but it pointed to a possible detriment of the therapy as circulating factors, such as creatinine, cystatin C, NT-proBNP, and endothelin-1, were all increased in those taking sildenafil.⁷³ A subanalysis was later performed on data collected from the RELAX trial to determine whether the lack of effect observed in that trial was due to a variation in patient population from the Guazzi study, which recruited a “high-pressure” cohort exhibiting more severe disease properties (hypertension, left ventricular hypertrophy, pulmonary hypertension, and RV dysfunction).⁸⁶ However, the authors determined that sildenafil actually worsened RV function in patients enrolled in the RELAX trial who began the study with RV dysfunction and impaired RV–pulmonary artery coupling.⁸⁶ One other study was published, which also recapitulated the findings of the RELAX trial; sildenafil (60 mg thrice daily for 12 weeks) offered no therapeutic benefit, but instead prevented reductions in pulmonary arterial wedge pressure and RV end diastolic pressure that occurred in the placebo group (NCT01726049).⁸⁷ Together, these earlier studies suggested PDE5i therapy is ineffective and potentially detrimental for patients with HFpEF.

Following the aforementioned studies, a meta-analysis was published, which determined PDE5i therapy was successful in studies whose patient populations had a majority prevalence of combined pre- and post-capillary pulmonary hypertension (Cpc-PH) concurrent with heart failure (heart failure with reduced ejection fraction trials and the Guazzi study).⁸⁸ On the other hand, the HFpEF studies, which showed no positive effect with sildenafil, recruited a patient population with a higher prevalence of isolated post-capillary pulmonary hypertension (Ipc-PH).^73,86–88 As a result of this meta-analysis, multiple studies have since been published investigating the use of PDE5i therapy in HFpEF patients with Cpc-PH. The first was a retrospective study, which reported that taking sildenafil (20 mg thrice daily) or tadalafil (40 mg once daily) for 12+ months improved RV function, exercise capacity, and overall clinical status.⁸⁹ This study was then followed by a small, open-label trial which showed sildenafil (25 mg thrice daily for 3 months, then 50 mg thrice daily for 3 months) improved pressures and function of the pulmonary circulation, right heart, and left heart, and subsequently improved patient exercise capacity.⁹⁰Although these 2 studies were promising, the most recent study, which was a large multicenter, double-blinded clinical trial deemed the PASSION trial, showed long-term use of tadalafil (20 mg once daily for 4 weeks, then 40 mg once daily for 5+ months) negatively impacted HFpEF patients with Cpc-PH. PDE5i therapy led to a higher frequency of serious adverse events and treatment discontinuation, as well as a higher risk for all-cause death relative to the placebo group.⁹¹ With such conflicting results, the question remains as to whether PDE5i can offer clinical benefit to patients with HFpEF, especially those with Cpc-PH.

Pitfalls of Phosphodiesterase 5 Inhibitor Trials

Many possible explanations underlie the conflicting results generated from PDE5i HFpEF clinical trials. One of the more obvious explanations determined in the later studies was that PDE5i therapy may only provide clinical benefit in a subset of patients—specifically those with concurrent Cpc-PH as opposed to those with Ipc-PH or no PH at all.⁸⁸ While Ipc-PH in HFpEF is driven from the backward propagation of increased left-sided pressures, Cpc-PH arises as a consequence of pulmonary vascular disease (in which NO bioavailability is reduced) in combination with the backward transmission of pressure from the left heart.⁹² Thus, the pathophysiology of Cpc-PH offers a direct target upon which PDE5 inhibition can act in patients with HFpEF. This has been successfully measured in 3 of the 4 studies examining PDE5i therapy in patients with HFpEF and Cpc-PH.^85,89,90 In fact, in 1 of these studies, it was shown that sildenafil specifically abolished only the pre-capillary component of PH, further suggesting that PDE5i is uniquely effective in resolving the vasculopathy occurring in Cpc-PH as opposed to the mechanical component also observed in Ipc-PH.⁹⁰ However, the most recent PASSION trial, which was a much larger, blinded study, reported a negative finding with tadalafil treatment for Cpc-PH-HFpEF.⁹¹ Although this study may indicate tadalafil should not be prescribed in this disease context, the possibility remains that the drug formulation itself, as opposed to the direct effect of PDE5i, drove the increase in adverse events and mortality risk. The PASSION study was terminated prematurely because of the discontinuation of drug distribution, which occurred due to potential contamination of the drug supply.⁹¹ Given patients had an increased risk of all-cause, not cardiovascular-related, mortality, it is possible that whatever contaminant drove the drug to be recalled reduced the overall safety of tadalafil.⁹¹

The varied response to PDE5i between patients with HFpEF and Cpc-PH and those with Ipc-PH may, in addition, be explained by PDE5 levels in these patients' hearts. Cardiac myocytes in a healthy heart only express PDE5 at minimal levels, if any, though it has been shown that PDE5 is upregulated in the myocytes of hypertrophied ventricles.⁹³ Perhaps during Cpc-PH-HFpEF, when pulmonary pressures are high enough to result in severe remodeling and dysfunction in the RV, PDE5 expression is increased to a level at which its inhibition could drive a measurable reversal of remodeling. A similar concept has been shown in a murine model of pressure overload, where sildenafil was only cardioprotective following severe, but not moderate, remodeling.⁹⁴ This may explain why sildenafil reduced RV size in patients with HFpEF and Cpc-PH but not Ipc-PH.^73,87,90 It is also possible that relieving the pre-capillary component of Cpc-PH reduces RV afterload, thereby reversing remodeling of the ventricle. Nonetheless, future studies should be conducted to expand upon the PDE5i trials in patients with HFpEF and Cpc-PH—a large, blinded trial is warranted that measures the long-term efficacy of PDE5i in improving cardiovascular function, exercise capacity, and long-term outcomes (ie, risk of hospitalization, mortality).

Another potential explanation for the neutral results in some of the HFpEF PDE5i trials is that PDE5 inhibition may have been insufficient at boosting cGMP levels. Among all the clinical trials described in this section, it remains to be seen whether treatment with PDE5i increased cGMP levels beyond what was observed in placebo. Most of the clinical trials performed did not measure circulating cGMP levels, though it is likely that the studies conducted in patients with HFpEF and Cpc-PH observed enough of an increase in cGMP to achieve therapeutic benefit. However, it is unclear if the studies that yielded neutral results only did so due to an inadequate increase in cGMP levels. The RELAX trial was the only study that measured cGMP levels, and while it reported increased circulating sildenafil levels in patients taking sildenafil, the increase in cGMP levels measured in that group was not statistically different from the change measured in the placebo group.⁷³ Therefore, it is impossible to identify whether sildenafil was ineffective in treating HFpEF in the RELAX trial because the drug itself was not achieving its function of increasing cGMP levels or because the drug's mechanism of action is generally insufficient for treating HFpEF.

In the interpretation of these studies, it is also worth considering that PDE5i therapy relies on the proper functioning of upstream nodes of the NO signaling pathway to achieve the largest therapeutic effect. If NO bioavailability is reduced and/or sGC becomes dysfunctional (as is the case in HFpEF), cGMP levels should be low at baseline without considering PDE5 activity. Preventing the degradation of pathologically low baseline levels of cGMP using PDE5i therapy should therefore only mediate minor therapeutic changes. Overcoming this potential problem would require targeting the NO signaling pathway at a second node (PDE5i + either nitrate, nitrite, or an sGC stimulator). Although targeting more than 1 node directly is likely to result in systemic hypotension, combining PDE5i therapy with an antioxidant, weight loss drugs, or exercise—any of which could lower oxidative stress to improve NO bioavailability and/or sGC function—may offer a way to increase cGMP production and induce a greater therapeutic benefit of PDE5 inhibition beyond what has already been reported.

PDE5i trials in patients with HFpEF may also be failing beyond treating Cpc-PH and RV dysfunction because not all of the cell types requiring targeting are being targeted by this therapeutic. For example, the PDE5i studies described here all reported no difference in systemic blood pressure, indicating the specificity of this therapy to cells of the pulmonary circulation.^73,85,88,90 Although improving pulmonary health is critical in HFpEF, the function of the systemic vasculature is equally important for its regulation of cardiac afterload and peripheral tissue perfusion. Like sGC stimulators, PDE5 inhibitors are also faced with the major limitation in that they only target muscle cells (vascular smooth muscle and cardiac myocytes) and not endothelial cells. As described in the previous section, without replenishing NO in endothelial cells, endothelial dysfunction is likely to persist, and HFpEF may remain difficult to resolve beyond the threshold measured in the successful studies described here.

It remains unclear what may have driven the negative outcomes observed in the PDE5i studies largely containing patients with HFpEF and Ipc-PH or no PH at all. In the RELAX trial, patients in the sildenafil group had increased biomarkers indicative of worsened renal function (eg, creatinine, cystatin C).⁷³ However, it has yet to be shown in other clinical trials that PDE5 inhibition worsens renal function. In fact, PDE5 inhibitors have actually shown beneficial effects on renal function in other disease contexts, including diabetic nephropathy and pulmonary arterial hypertension.^95,96 A more recent analysis, which calculated RELAX trial subjects' estimated glomerular filtration rates (eGFR) across multiple timepoints (baseline, 3, 12, and 24 weeks of treatment) using the 2021 Chronic Kidney Disease Epidemiology Collaboration definition,⁹⁷ showed sildenafil did not affect the eGFR slope (the change in eGFR over time).⁹⁸ This study also determined that plasma levels of sildenafil were actually negatively correlated with eGFR decline between 0 and 12 weeks of treatment, though this correlation was lost by the 24-week end point. Based on these results, it is possible that PDE5i therapy may not worsen kidney function. However, a follow-up study would be warranted to validate this conclusion.

In the other trial reporting a negative outcome with PDE5i therapy, sildenafil prevented improvements in pulmonary arterial wedge pressure and RV end diastolic pressure that were observed in placebo.⁸⁷ This could potentially be due to the imbalance in patient profiles between the placebo and sildenafil groups. The placebo group tended to have a higher percentage of subjects taking diuretics, angiotensin converting enzyme inhibitors/angiotensin receptor blockers, and mineralocorticoid receptor antagonists, likely due to these patients also having a higher prevalence of comorbidities, including hypertension. Perhaps this varied use of medications, which relieve high blood pressure and volume overload, mediated the differences achieved. However, this is unlikely, as the authors reported there being no change in the dosage of these medications nor of patient weight throughout the duration of this study to account for the differences between the treatment groups.⁸⁷ This leaves the possibilities that either 1 (or more) of the underlying comorbidities more prevalent in the placebo group (ie, coronary artery disease, diabetes, hypertension, atrial fibrillation) somehow mediated this improvement, which is improbable, or that this result occurred by random chance, which was noted by the authors.⁸⁷ A follow-up study with more balanced patient profiles between treatment groups may help to determine whether PDE5i therapy is actually worsening patient right and left heart pressures.

Future Considerations

As described throughout this review, the success of therapeutics targeting the NO signaling pathway in patients with HFpEF has remained variable. Each category of therapeutics—nitrates, nitrites, sGC stimulators, and PDE5 inhibitors—harbors its own set of limitations and pitfalls that need to be addressed to observe consistent success (Fig. 2). One factor that must continue to be considered in the design of HFpEF trials is the recruited patient population. In inorganic nitrate studies, female patients showed limited benefit. Meanwhile, nitrite and PDE5i studies demonstrated that those with PH-HFpEF (specifically Cpc-PH for PDE5i) experienced therapeutic benefit, while studies with broader inclusion criteria did not achieve similar outcomes. HFpEF is known for being a heterogeneous condition; therefore, it is unlikely there will be a “one-size-fits-all” therapy. Future studies should aim to target HFpEF subpopulations that have achieved positive outcomes in the smaller trials described. Alternatively, larger trials on the efficacy of these therapeutics should be conducted to allow for subgroup analyses that may uncover a patient population uniquely benefiting from treatment.

FIGURE 2.

Pitfalls of therapies targeted to nitric oxide signaling in heart failure with preserved ejection fraction. NO signaling–based therapies have shown inconsistent impacts on patients with HFpEF. The deleterious outcomes of treating patients with HFpEF with organic nitrates have been attributed to nitrate tolerance and associated reactive oxygen–nitrogen species (RONS). Inorganic nitrates have conversely shown mixed efficacy, with the neutral outcomes potentially being due to impaired drug conversion from nitrate to nitrite, therapeutic dosing, and/or sex of the patient population. Inconsistent efficacy of nitrite therapy may be explained by impaired drug delivery, comorbidity status of the patient population (more specifically, whether the patients are recruited based on the presence of pulmonary hypertension), and/or the short half-life (30–40 minutes) of nitrite therapy. Soluble guanylyl cyclase (sGC) stimulators may be ineffective due to the heme status of sGC. sGC stimulators require ferrous heme to be functional, though, with oxidative stress (like in HFpEF), sGC heme can become oxidized or lost, preventing drug activity. sGC stimulators and phosphodiesterase 5 (PDE5) inhibitors may also not be more effective as they only target cardiac and vascular muscle cells, but not endothelial cells. Furthermore, PDE5 inhibitors may only show minimal benefit as a result of comorbidity status (specifically, the presence or absence of combined pre- and post-capillary [Cpc] pulmonary hypertension), inability to prevent cyclic guanosine monophosphate (cGMP) breakdown, and reliance on upstream nodes of the NO signaling pathway, which have already been shown to be dysfunctional during disease. All of the aforementioned drugs are also sensitive to the oxidative environment in HFpEF, which may be preventing NO supplementation, sGC stimulation, and subsequent increases in the cGMP pool, which PDE5 inhibitors work to maintain. Figure produced using BioRender.com.

Oxidative stress is another factor that likely must be addressed to observe greater success from NO-based therapies in HFpEF. Systemic oxidative stress is thought to precede reduced NO bioavailability in HFpEF.⁸ As potent NO scavengers, RONS (primarily superoxide) are likely impeding NO supplementation through nitrate or nitrite as well as negatively impacting sGC heme status to both dampen sGC function and response to stimulation. Action should therefore be taken to reduce oxidative stress while simultaneously replenishing NO signaling. Directly targeting RONS using antioxidant treatment in patients with HFpEF was sufficient to independently increase nitrite levels, as well as flow-mediated dilation.⁹⁹ Thus, antioxidant therapy in combination with inorganic nitrate/nitrite may further amplify this response. As superoxide and peroxynitrite are key contributors to both the oxidative environment and the reduction in NO bioavailability during HFpEF, antioxidant therapies targeting these RONS may offer the optimal outcome. Another way to target RONS indirectly is through the treatment of cardiometabolic comorbidities (eg, obesity, type II diabetes), which promote the stressed state. Pairing exercise with NO-based therapies is thus an intriguing avenue worth examination. The few studies pairing exercise training with NO-based therapies had neutral outcomes, though these studies were hampered by dosing and timing issues, which may have prevented any benefit.^56,69 With the rise of drugs targeting obesity and/or type II diabetes (eg, SGLT2 inhibitors or GLP-1 agonists) in HFpEF, combining the use of either of these drug classes with NO-based therapies may prove beneficial. Currently underway is the phase II SAK HFpEF trial (NCT05138575), which is investigating how empagliflozin, with or without KNO₃, impacts exercise capacity. Although this study will not randomize any patients to receive KNO₃ without empagliflozin, it will offer great insight into whether a synergistic effect occurs in combining the 2 therapies. The group conducting this trial is concurrently recruiting for another trial examining the use of KNO₃ (to increase tissue perfusion) in combination with propionyl-l-carnitine and nicotinamide riboside (supplements that may improve mitochondrial function) to improve exercise capacity in patients with HFpEF. This study, deemed the MPMA trial (NCT04913805), will provide further insight as to the efficacy of inorganic nitrates in altering exercise capacity on its own as well as in a dual therapy.

CONCLUSIONS

Therapeutics targeting the NO signaling pathway have yielded largely neutral or contradictory results in human HFpEF trials despite the success observed in animal models of the disease. Although the lack of consistent success may suggest that NO therapies should not be indicated in the treatment of HFpEF, many of these trials presented with potential problems in dosing, duration, drug delivery and metabolism, and/or the patient population recruited. These trials were further limited by the likely presence of RONS (from comorbidities in patients with HFpEF and/or chronic organic nitrate use), which could prevent efficient NO donation and sGC stimulation. Negative results occurred consistently in organic nitrate trials, potentially because of vascular induction of endothelin-1 and hypersensitization to the neurohormonal environment. Based on this consistency in negative outcomes, future efforts targeting the NO signaling pathway in HFpEF should focus on the other therapeutics described. Excluding organic nitrates, it remains difficult to definitively conclude that NO-based therapies are not sufficient to treat HFpEF. Although nitrates, nitrites, sGC stimulators, and PDE5 inhibitors are not currently approved for use in HFpEF, future studies may reveal that, under the right conditions, targeting the NO signaling pathway (independently or in combination with another therapeutic) could provide clinical benefit in patients. However, further investigation is required to determine this.