OUP user menu

Antioxidant treatment for heart failure: friend or foe?

M. J. Thomson, M. P. Frenneaux, J. C. Kaski
DOI: http://dx.doi.org/10.1093/qjmed/hcn160 305-310 First published online: 17 December 2008


Increasing studies demonstrate a pivotal role for oxidant stress in the pathophysiology of heart failure (HF). Recent meta-analyses also reveal the potential pitfall of a mono-dimensional antioxidant approach. This review article summarizes the main biological pathways involved in oxidant stress and HF, the possible deleterious nature of certain antioxidant monotherapy and proposes potential antioxidant strategies necessary to challenge specific HF aetiology and progression.

Oxidant stress is well established in the aetiology of cardiovascular disease (CVD) and increasingly implicated in congestive heart failure (CHF). Human and animal models of CHF demonstrate enhanced reactive oxygen species (ROS) production, reduced antioxidant reserve and ROS-mediated cardiac injury. Similarly, various ‘antioxidant’ agents used in the treatment of CHF appear to confer cardio-protective benefits. In contrast, recent meta-analyses have suggested a deleterious effect of vitamin E in heart failure (HF), warranting reappraisal of ‘antioxidant’ use in this context.

HF and oxidant stress

CHF is characterized by chronic and progressive left ventricular (LV) systolic dysfunction associated with neurohormonal dysfunction and the chronic activation of the renin-angiotensin-aldosterone system (RAAS); a current pharmacological approach to HF involves suppressing the neurohormonal interplay between RAAS and ROS production.1

Chronic HF is associated with aberrant oxygen-derived free radical production, such as superoxide anion generation that promote nitric oxide (NO) biodegradation; the primary mediator regulating vascular homeostasis.2 HF thereby promotes endothelial dysfunction, which is characterized by impaired flow-dependent endothelial-dependent vasodilation and dysregulation of tissue perfusion, thus causes a vicious cycle of ROS production.1 Numerous enzymatic sources are implicated in ROS generation in HF including: xanthine oxidase/oxidoreductase (XO/XOR),3 vascular and phagocyte nicotinamide adenine dinucleotide phosphate (NADPH) oxidase,4 auto-oxidation of catecholamines,5 uncoupling of NO synthase (NOS),6 infiltrating inflammatory cells7 and mitochondrial leakage.8,9 ROS adversely affect cardiomyocyte function via: inducing cellular injury and apoptosis, depressing myocardial contractility and alteration of the cardiovascular systems redox equilibrium through dysregulated signal transduction, transcription and gene expression.9–13

Endogenous ‘antioxidants’ and HF

Aberrant ROS production is countered by endogenous antioxidant defences including antioxidant vitamins, such as vitamin C and vitamin E and antioxidant enzymes like superoxide dismutase (SOD) and glutathione peroxidase (GP), located in specific cellular compartments.9 Human arteries contain enhanced SOD compared with other tissues; vascular SOD is reported to protect NO from the antagonistic actions of superoxide ions. Landmesser et al.1 have shown that CHF patients with idiopathic dilated cardiomyopathy have significantly enhanced endothelial XO activity and reduced SOD, compared with controls. Patients with low SOD and enhanced XO levels showed the greatest improvement with the administration of vitamin C regarding restoration of flow-dependent endothelial-mediated vasodilation. This finding suggests that subgroups of CHF patients with enhanced oxidant stress and reduced oxidant reserve may benefit from antioxidant treatment.1

Further, in a mouse model of HF caused by myocardial infarction (MI), over-expression of GP significantly attenuated LV remodelling, reduced myocyte hypertrophy, apoptosis, fibrosis, LV dilatation and dysfunction, and improved survival.14 In addition, Adamy et al.15 demonstrated that LV glutathione, an intracellular antioxidant, was deficient in a rat model of CHF post-MI compared with control animals. Levels of glutathione, however, normalized within 1 month of the oral glutathione precursor N-acetylcysteine (NAC). Restoration of glutathione was associated with improved LV function and reduced LV remodelling. Interestingly, glutathione repletion correlated with the inhibition of key pro-inflammatory pathways involving tumour necrosis factor alpha (TNF-α), known to be implicated in CHF progression.16 These finding suggest that NAC and antioxidant enzyme precursors may prove a novel potentially effective therapy in CHF, but whether animal models can be directly extrapolated to human CHF remains speculative.

Under physiological circumstances, vitamin C is concentrated up to 10-fold within cells, and acts principally as an intracellular antioxidant where it inhibits low-density lipoprotein oxidation, reduces oxidized glutathione and functions as a co-antioxidant to vitamin E.17,18 Vitamin C uptake is an active insulin-dependent process, and there is an association between enhanced myocardial insulin resistance and HF, suggesting that vitamin C uptake may be reduced in HF.19,20 Vitamin C is also known to enhance NO production via chemical stabilization of tetrahydrobiopterin (BH4), a co-factor in NO production, thereby promoting vascular homeostasis.21 When administered at high doses, vitamin C enhances endothelial cell proliferation, and stabilizes and enhances collagen—and other extracellular matrix protein formation—potentially contributing to restore vascular integrity and to improve re-modelling.22,23 In a placebo controlled trial involving 33 patients with chronic HF, Piccirillo et al.24showed that treatment with high-dose vitamin C significantly enhanced baroreceptor sensitivity, a variable reduced in HF. Further clinical trials in larger numbers of patients, however, will be required to substantiate this potentially important observation and determine if the effects are sustained over longer periods of time. Along the same lines, Nightingale et al.25demonstrated, in a placebo controlled trial in 100 patients with CHF, that acute intravenous vitamin C (single 2 g dose), but not chronic oral vitamin C (4 g/day for 4/52) administration improved baroreceptor sensitivity, perhaps indicating that oral administration of vitamin C may provide inadequate free radical scavenging ability; an important suggestion that requires further investigation.

Epidemiological data and some clinical trials suggest cardioprotective benefits from certain diets rich in antioxidants, such as the Mediterranean diet. The Lyon Diet Heart Study assessed 605 patient survivors of a MI who were randomized to receive a ‘Mediterranean style’ diet or a ‘control’ diet. Patients in the Mediterranean diet group showed a significantly reduced (70%) risk of serious cardiovascular events, i.e. unstable angina, recurrent MI and death.26,27 Intriguingly, large clinical trials using antioxidant vitamins have not achieved such impressive results. These controversial findings have challenged the usefulness of the ‘antioxidant’ approach to the management of atherosclerosis and HF.28–31 These findings suggest that antioxidant vitamins used as a mono-dimensional antioxidant strategy are inadequate and may reflect insufficient bioavailability and potency of oral antioxidant vitamins compared with, for example, a varied diet rich in a range of antioxidants and potential co-factors; this requires further research. Figure 1 outlines the main biochemical players involved in the delicate balance between antioxidant reserve and oxidant stress in the development of HF.

Figure 1.

A summary of the main inhibitors and promoters of HF: oxidant stress versus antioxidant reserve.

HF and NO

Recently, the significance of an altered nitroso–redox state in HF has come to the fore. NO is implicated in regulating heart and vascular function.32 Experimental models demonstrate attenuated NO production in HF, and NO is known to inhibit XOR and NADPH oxidase.9,33 NO plays a pivotal role in cellular signalling through post-translational modification of effector molecules via S-nitrosylation of cysteine residues. Oxidative stress disrupts NO signalling by two principal means: (i) interference of S-nitrosylation and (ii) direct reaction with NO resulting in the production of peroxynitrite, a highly reactive and deleterious intermediate agent. S-nitrosylation governs ion channel function in the myocardium, responsible for systolic and diastolic function, modulates conductance vessel resistance and, in the coronary microcirculation, also modulates blood flow responses.32 S-nitroso-haemoglobin is reduced in HF and carries NO to the microvasculature; it has been suggested that it may function as an endocrine vasomodulator when local oxygen tension (PO2) is reduced, as in HF patients.34,35

NOS catalyses NO production and is spatially and temporally regulated.33 The heart contains at least three NOS isoforms, which are discretely compartmentalized: (i) neuronal NOS (nNOS), situated in the sarcoplasmic reticulum and involved in both neurotransmission and β-adrenergic stimulation; (ii) constitutively formed endothelial NOS (eNOS), generated from endothelial cells, the endocardium and cardiomyocytes and (iii) inducible NOS (iNOS), produced by activated cardiomyocytes and macrophages.36–39 The latter two isoforms are implicated in governing myocardial contractility.37 Interestingly, post-MI mouse models with nNOS homozygous deletions have an enhanced XOR activity, reduced LV function and increased mortality. This suggests that NO may have cardioprotective effects in LV failure post-MI, despite the potential contribution of nNOS to β-adrenergic hyporesponsiveness.33

Human in vivo models also support the theory that NO modulates the Frank–Starling responses. In patients with dilated cardiomyopathy iNOS/eNOS, up-regulation linearly correlates with augmented stroke volume and a significantly enhanced LV stroke work.40 In contrast, it has been suggested that certain NOS isotypes, in the absence of counter-regulation, can promote nitrosative stress in the HF setting. In some studies, whilst iNOS production has been associated with cardiac decompensation, eNOS down-regulates iNOS thus resulting in both improved endothelial and LV function.37,41 Thus, the heterogeneity of NOS responses in HF needs to be investigated further to elucidate the differential regulation and effects of NOS isoforms and nitroso–redox balance.39

Clinical ‘antioxidants’ and HF

Angiotensin-converting enzyme inhibitors (ACE-), β-blockers and aldosterone antagonists have proven clinical benefit in HF, and their pharmacological actions include potent ‘antioxidant’ properties. However, despite their usefulness in the treatment of post-MI patients and patients with HF, morbidity and mortality remains high among CHF patients.7,9 Angiotensin II inhibition and β-blockade enhance eNOS expression; similarly, allopurinol, a XOR inhibitor, significantly improves myocardial efficiency in idiopathic dilated cardiomyopathy, suggesting that XO may be a significant source for ROS in the vasculature.1,42 Studies in mouse models of cardiac fibrosis, with a deficiency in the catalytic subunit Nox2 of NADPH oxidase, demonstrated that Angiotensin II mediates Nox2-NADPH-dependent interstitial cardiac fibrosis and this is inhibited by the mineralocorticoid inhibitor spironolactone. This finding suggests that mineralocorticoid receptor activation has important downstream effects that promote the transition between RAAS activation, cardiac fibrosis and CHF. The precise molecular mechanisms responsible for these effects remain to be determined.43

Along these lines, a randomized controlled trial in black patients demonstrated that combined hydralazine and isosorbide mononitrate treatment significantly improves HF outcome and slows HF progression. It is conceivable that as these pharmacological agents augment NO bioavailability—isosorbide mononitrate is a NO donor and hydralazine possesses ‘antioxidant’ properties by inactivating ROS generated by NADPH oxidase—they can confer protection by avoiding or slowing down NO degradation.44 Figure 2 summarizes the main proposed mechanism of action of some ‘antioxidant’ agents in HF in the disequilibrium between oxidant stress and antioxidant reserve.

Figure 2.

A summary of some of the proposed mechanisms of action of various ‘antioxidant’ agents in the delicate interplay between oxidant stress and antioxidant reserve in HF.

Is vitamin E deleterious in HF?

Trials of supplementary antioxidant vitamin therapy have failed to produce clinically relevant beneficial effects. The HOPE-TOO trial45 reported that patients with vascular disease or diabetes mellitus prescribed 400IU vitamin E had a higher risk of HF (P = 0.03) and an increased number of hospitalizations (P = 0.045). Similarly, in the GISSI-Prevenzione study29 vitamin E treatment was associated with a 50% increase in the development of HF. Furthermore, a recent meta-analysis showed that high-dosage vitamin E administration (>400 IU) was associated with an increased risk of all-cause mortality in a dose-dependent fashion (P = 0.035).46–48 These findings have led some authors to caution against the use of supplementary vitamin E.45 Vitamin E has been suggested to have pro-oxidant effects in the absence of co-antioxidants, and to fail to augment antioxidant defences in vivo.49 Moreover, vitamin E has no effect on certain ROS species, such as hypochlorite-induced oxidation, interferes with lipoprotein metabolism, and diminishes high-density lipoprotein-2 cardioprotective effects. In contrast, in CHF vitamin C suppresses endothelial cell apoptosis and improves endothelial function.50,51 In a recent study, however, we showed that 1 month of vitamin C therapy had no effect on exercise capacity (peak V02) and worsened skeletal muscle energetic status, despite improving endothelial function in patients with HF.52

Summary and conclusions

There is compelling evidence that ROS-mediated cardiac injury is implicated in the development of CHF. Numerous clinically proven pharmacological agents in HF have cardioprotective benefits, beyond their designated pharmacological role, including ‘antioxidant’ properties. Despite this, ‘antioxidant’ vitamin trials in HF, and also in atherosclerotic heart disease, have been disappointing so far, and a meta-analysis concluded that vitamin E, given as monotherapy in certain patient categories, may contribute to the development or impairment of HF.

ROS dysregulate signal transduction and transcription from cardiomyocytes to endothelium, and disturb the nitroso–redox equilibrium.51 Thus ‘antioxidant treatment’ directed to counteract pivotal ROS-activated pathways appears as a rationale intervention in HF, as the inhibition of oxidant stress may favourably modulate myocardial inotropic function and Frank–Starling curve responses. This however, may not be straightforward as specific ROS may have to be tackled in different HF aetiologies, and the efficacy of the intervention is likely to depend on the complex responses that it may elicit. The intervention is also likely to affect the complex interaction among myocytes, endothelial cells and neurohormonal pathways in CHF, making it difficult to predict their net effects. However, studies such as the Lyon Diet Heart Study suggest promise in the use of a multifaceted antioxidant approach through, for example, an antioxidant diet and imply that agents with varied ‘antioxidant’ properties may prove most beneficial.

Investigation of atherogenesis in recent years has incorporated markers of oxidant stress and antioxidant reserve, using methods such as mass spectrometry to identify specific molecular footprints of oxidative stress and nanosensing techniques to assess NO production.53,54 A similar approach may be required to divulge oxidant burden and antioxidant reserve, as well as identifying the most appropriate ‘antioxidant’ regime and temporal targeting, in a given patient with HF. We already know some of the triggers to CHF, but halting the specific deleterious pathways that are activated in this setting and elucidating which ‘antioxidants’ efficaciously halt these differential effects, represents a major challenge confronting basic science researchers and clinicians alike.

Conflicts of interest: None declared.


View Abstract