Alterations in the Sympathetic and Parasympathetic Nervous Systems in Heart Failure

Catecholamines

Venous plasma NE concentrations, acquired at rest, reflect primarily neurotransmitter release from upstream forearm skeletal muscle and provide little insight into the magnitude and duration of other sympathetic nerve or adrenal responses to emotional stimuli or exercise. Moreover, the effect of heart failure or its therapies on neurotransmitter release cannot be deduced definitively solely from changes in plasma catecholamine concentrations. The majority of NE released from sympathetic vesicles is subject to neuronal or extraneuronal uptake; only a small fraction either acts on postjunctional adrenoceptors or appears in plasma. Low cardiac output reduces neuronal and extraneuronal clearance, causing plasma NE concentrations to rise.

The isotopic method addresses such limitations but at increased complexity and cost. Total body NE spillover into plasma is determined from the dilution of tritium-labeled NE during its steady-state infusion in tracer concentrations. If venous effluent from an organ, such as the heart, kidney, or brain, is collected simultaneously with an arterial sample, the difference in tritium-labeled NE between vein and artery can be used to calculate its local extraction (Extr) by neuronal and extraneuronal transport mechanisms. Organ-specific NE spillover (NES) can then be determined by the equation:

with NE _a and NE _v representing arterial and venous concentrations of unlabeled NE and PF plasma flow. Neuronal NE re-uptake and storage can be estimated by calculating spillover of tritium-labeled dihydroxyphenylglycol (DHPG), an intraneuronal metabolite.

Microneurography

Multifiber or single-unit recordings from postganglionic nerves supplying muscular or cutaneous vascular beds capture the dynamic nature of sympathetic nerve firing and illustrate its reflex control ( Fig. 13.1 ). Skin sympathetic nerve activity is not modulated by input from high- or low-pressure baroreceptor reflexes. It responds preferentially to alerting stimuli or cold with burst firing occurrence independent of the cardiac cycle, whereas sympathetic discharge directed at skeletal muscle resistance vessels is entrained by input from arterial and cardiopulmonary mechanoreceptors. Muscle sympathetic nerve activity (MSNA) exhibits distinctive pulse-synchronicity, with multiunit bursts appearing 1.1 to 1.3 seconds after the preceding R wave of the electrocardiogram. In healthy subjects, MSNA is activated by reductions in diastolic or cardiac filling pressure, exercise, hypoxia, hypercapnia, and arousal from sleep, and is inhibited by lung inflation. In healthy resting individuals, the magnitude of MSNA correlates with both renal and cardiac NE spillover. Isometric exercise elicits proportionately similar increases in MSNA and cardiac norepinephrine spillover (CNES). In HFrEF, however, such concordance is lost.

Arterial Baroreflex Sensitivity

Because cardiac cycle length responds more rapidly to the release and metabolism of acetylcholine than of NE, the normally brisk sinus node response to an acute perturbation in blood pressure is mediated primarily by arterial baroreceptor reflex-mediated vagal activation or withdrawal. One method of determining the strength, or gain, of reflex vagal heart rate regulation is to quantify the bradycardic response to a bolus of a vasoconstrictor drug, such as phenylephrine, or the tachycardic response to nitroprusside- or nitroglycerin-induced hypotension. Longer (“steady state”) infusions of these vasoactive agents elicit competing sympathetic influence. The use of such drugs for this purpose has several limitations. Nitrate donors affect sino-atrial discharge directly. By causing sustained vasoconstriction or dilation, these drugs mechanically distort baroreceptor nerve endings. Their action is nonspecific, affecting also mechanoreceptors situated in the atria and pulmonary vasculature. Heart rate responses to carotid sinus baroreceptor stimulation by neck suction or unloading by neck pressure have been studied, but aortic arch baroreceptors will elicit counter-regulatory reflex responses to any arterial pressure changes these maneuvers induce.

Algorithms have been developed to track spontaneous fluctuations in blood pressure and heart rate from continuous noninvasive or invasive recordings, and to identify, within such time periods, brief sequences with concordant changes in systolic blood pressure and the subsequent R-R intervals (inverse of heart rate).

Common to such methods is the construction of regression equations relating changes in output (pulse interval; in milliseconds) to changes in input (systolic blood pressure; mm Hg) from antecedent cardiac cycles. The slope of this estimates the gain of the arterial baroreflex control of heart rate. Values obtained by this spontaneous sequence method are qualitatively similar to those derived using vasoactive drugs.

Reflex regulation of central sympathetic outflow can be evaluated similarly by relating changes in MSNA to changes in arterial pressure during graded infusions of pressors and vasodilators, or by constructing, under resting drug-free conditions, regression equations relating spontaneous changes in MSNA to preceding diastolic blood pressures. By their nature, such methods cannot establish if any reduction in the arterial baroreflex gain of either heart rate or sympathetic nerve firing accrues from conduit artery inelasticity, diminished neural transduction of baroreceptor stretch, central neuroplastic adaptation, or efferent nerve dysfunction.

Heart Rate Variability

The half-life of acetylcholine is too brief and its actions too local to permit assay. Instead, tonic vagal heart rate modulation is estimated indirectly, most often and best validated by determining beat-to-beat variation either within the time domain (e.g., the standard deviation of all nonectopic pulse intervals occurring within a specified period) or within the frequency domain, using spectral analysis. A complimentary method, termed heart rate turbulence analysis, evaluates beat-to-beat variations in heart rate that follow a premature ventricular complex. Algorithms commonly employed to derive heart rate power spectra include Fast Fourier Transformation, autoregression, and coarse-graining spectral analysis (CGSA). The complexity of the heart rate variation is such that a range of additional analytic methods, both linear and nonlinear, have been proposed to characterize the information contained more fully, but few have been adopted broadly for research, clinical, or prognostic purposes.

Because atropine abolishes high-frequency (0.15–0.5 Hz) spectral power, oscillations in heart rate within this band have been attributed to parasympathetic activity, with respiration as its primary rhythmic stimulus. Conversely, because maneuvers known to increase central sympathetic outflow, such as standing, tilt, and exercise, increase low-frequency (0.05–0.15 Hz) spectral power, whereas decreases are observed during sleep or after β-blockade or central sympatholysis with clonidine, heart rate fluctuations within these frequencies initially were considered specific representations of sympathetic neural modulation. However, parasympathetic oscillatory input also influences power spectral frequencies below 0.15 Hz. For this reason, a (contentious) ratio between low- and high-frequency power has been proposed as an estimate of “sympathovagal balance.”

Oscillatory autonomic contributions to heart rate variability are superimposed on a broadband nonharmonic fractal signal, most prominent between 0.00003 and 0.1 Hz (i.e., at very low and low frequencies). Using CGSA, this nonharmonic power can be quantified by plotting the log of spectral power as a function of the log of frequency (l/f ^β plot) then extracted, yielding more precise estimates of residual harmonic contributions to low- and high-frequency (0.15–0.50 Hz) power. For this reason, CGSA becomes particularly useful when evaluating autonomic contributions to heart rate variability in patients with heart failure, whose harmonic spectral power is both concentrated within the very low- and low-frequency ranges and are markedly diminished relative to the nonharmonic signal.

Frequency domain analysis should be appreciated primarily for the insight it allows into mechanisms responsible for short- and long-term heart rate oscillations and for its prognostic value in populations with cardiovascular disease. At best, it provides an estimate of the extent to which parasympathetic and cardiac sympathetic discharge and neurotransmitter release modulate heart rate within these specific frequency bands, but not of the intensity of such neural discharge or the magnitude of sympathetic outflow directed elsewhere, for example, to the ventricle, kidney, or regional vascular beds.

Cross-Spectral Analysis

Algorithms used to derive power spectra for heart rate can also be applied to blood pressure, respiratory signals, and MSNA. Cross-spectral analysis between two such variables can establish their coherence, the influence of input on output, and the phase delay between these signals. For example, the gain of the transfer function between systolic blood pressure (input) and pulse interval (output) oscillations within the low- or high-frequency regions (α coefficient) estimates arterial baroreceptor reflex control of heart rate (although agreement with values obtained using vasoactive drug methods is poor), and the gain of the transfer function between blood pressure (input) and MSNA (output) estimates arterial baroreflex modulation of central sympathetic outflow.

Tracer Imaging With Catecholamine Analogues

Nuclear iodine-123 metaiodobenzylguanidine ( ¹²³ I-MIBG) or ¹¹ C-metahydroxyephredrine ( ¹¹ C-HED) positron emission tomographic (PET) imaging infer the integrity and homogeneity of cardiac sympathetic innervation and NE transport and indirectly, by calculating tracer uptake and washout, an approximation of its neural release. PET imaging provides superior spatial resolution. PET tracer analogues of β-adreno-, muscarinic, and nicotinic receptors are being evaluated for potential clinical application.

Heart Failure With Reduced Systolic Function

The onset of HFrEF is characterized by a cardiac-specific autonomic “signature,” with the loss of tonic and reflex vagal heart rate modulation and selective increase in cardiac NE spillover; total body or renal NE spillover and MSNA are not elevated. Indeed, in many patients with asymptomatic or mild to moderate symptomatic HFrEF, plasma NE concentrations and sympathetic nerve firing rates are similar to those of age-matched healthy subjects ( Figs. 13.2 and 13.3 ).

Sympathetic Activation

Relative to the unaffected population, cohort-mean plasma NE concentration increases as asymptomatic HFrEF progresses to overt congestion. In more advanced HFrEF, heightened adrenal sympathetic nerve activity stimulates medullary epinephrine and NE release into plasma.

In health, approximately 25% of total body NE spillover arises from the kidney and about 2% from the heart. In patients with left ventricular dysfunction and congestion, studied before the advent of contemporary therapies, calculated NE clearance was one-third lower, and total body NE spillover double that of control subjects. Approximately 60% of this increase resulted from a 5- to 20-fold elevation in cardiac and a 2- to 3-fold greater renal NE spillover into plasma. Mental stress and cycling exercise elicit further increases in cardiac neurotransmitter release, indicating preservation of myocardial adrenergic reserve. Preferential activation of cardiac sympathetic nerve traffic also is evident in the paced-ovine HFrEF model.

Microneurographic recordings provided definitive evidence that the majority of this increase in total body NE spillover can be attributed to greater central sympathetic outflow rather than altered neurotransmitter reuptake or clearance. Single-unit recordings from patients with HFrEF in sinus rhythm display both higher firing probability and recruitment of previously silent efferent neurons, but no increase in the proportion of cardiac cycles associated with multiple discharges; whereas those who develop atrial fibrillation exhibit a significantly greater incidence of multiple firing, which would evoke more NE release and consequently greater vasoconstriction. In contrast, sympathetic discharge to skin is not increased.

There is little or no correlation, in patients with HFrEF, between left ventricular ejection fraction (LVEF) and MSNA burst incidence. Indeed, if oxygen uptake at maximum exercise capacity (peak VO ₂ )is relatively preserved, resting MSNA can remain within the range of age-matched control subjects, despite profound left ventricular systolic dysfunction. However, their sympathoneural response to exercise is exaggerated ( Fig. 13.4 ).

In the frequency domain, heart failure and age-matched control subjects display similar total MSNA power, harmonic power, nonharmonic power between 0 and 0.5 Hz, and spectral density within the very low (0–0.05 Hz) and high (0.15–0.5 Hz) frequency bands. However, low-frequency oscillations in the mean voltage neurogram are diminished markedly or absent, despite near-maximal sympathetic burst incidence, indicating loss of central or reflex modulation of efferent sympathetic traffic proportional to heart failure severity.

Heart Failure With Preserved Ejection Fraction (see also Chapter 11 , Chapter 39 )

Our understanding of this phenotype would benefit from more extensive and rigorous investigation. To date, HFpEF cohort sizes have been small, the definitions of HFpEF inconsistent (some conflate hypertension plus impaired relaxation with HFpEF), the quality of studies published often have been suboptimal, and few have considered the confounding excitatory influences of age and left ventricular hypertrophy, with blunted stimulation of inhibitory ventricular afferents, on sympathetic activity. Although long-standing hypertension is a well-recognized risk factor for HFpEF and itself is often accompanied by chronic sympathetic excitation, no prospective studies have as yet determined whether such long-term autonomic imbalance increases the risk of incident HFrEF.

In the Studies of Left Ventricular Dysfunction (SOLVD) registry, mean plasma NE concentrations of patients with pulmonary congestion and LVEF greater than 45% were not increased. In general, heart rate variability is diminished, the sympathetic nervous system appears less activated than in patients with HFrEF, and cardiac NE transport function, as assessed using ¹²³ I-MIBG imaging, not as defective. A series of 15 patients with restrictive cardiomyopathy due to endomyocardial fibrosis and a LVEF greater than 50 who were studied at least 6 months after endocardial resection surgery described increases in MSNA burst frequency and incidence, sympathetic contributions to heart rate variation, and decreases in indices of tonic and reflex vagal heart rate modulation that were comparable to those observed concurrently in a matched cohort of patients with dilated cardiomyopathy. The recent discovery of an excitatory reflex engaged by high atrial pressure (vide infra) identifies a potential causal mechanism that could elicit sympathoexcitation in both HFrEF and HFpEF.

Cardiac

Fractional shortening is diminished in mice with defective vesicular acetylcholine transport; deficient neurotransmitter release is accompanied by more extensive fibrosis in response to exogenous angiotensin II. The impact of impaired vagal tone in human heart failure on heart rate modulation, inflammatory pathways, the regulation of left ventricular performance, and the progression of HFrEF have been reviewed in detail.

The combination of early vagal withdrawal plus a selective increase in cardiac NE spillover constitute adaptive autonomic responses, which are engaged to maintain peripheral tissue perfusion in the face of compromised ventricular performance. However, once congestion becomes manifest, the heart is subject to the greatest proportional increase in regional NE spillover, and thus, the failing heart is the organ exposed for the longest duration to the greatest magnitude of sympathetic activation. The direct adverse myocardial consequences of such intense cardiac adrenergic drive, which are reviewed in Chapter 6, Chapter 10 , include: an increase in heart rate; induction of myocyte necrosis and apoptosis; fibrosis; decreased β ₁ -adrenergic receptor density; diminished β ₁ -adrenoceptor responsiveness to catecholamines and altered β ₁ -adrenergic receptor signal transduction with upregulation of G-protein-coupled receptor kinase 2; defective calcium regulation by the sarcoplasmic reticulum; induction of proinflammatory cytokine expression, greater oxidative stress; increased muscarinic M2 receptor expression; destruction of sympathetic nerve terminals; and depletion of myocardial NE content—all of which contribute to the relentless progression of cardiac dysfunction. Nonuniform NE depletion and sympathetic denervation disturb the temporal coordination of right and left ventricular contraction and relaxation and alter the dispersion of refractoriness, leading to ventricular dyssynergy and promoting arrhythmogenesis. In human experiments, muscarinic stimulation was found to exert a negative lusitropic effect and to antagonize the effect of beta-adrenergic stimulation in patients with HFrEF.

Peripheral

Sympathetically mediated constriction of capacitance and resistance vessels increases both preload and afterload; diminished conduit artery compliance impairs ventricular-vascular coupling. Stimulation of renal sympathetic nerves activates the renin-angiotensin-aldosterone axis ( see also Chapter 15 ), promotes tubular absorption of sodium and water, increases renal vascular resistance, and blunts the renal responsiveness to atrial natriuretic peptide. Angiotensin II–mediated efferent arteriolar vasoconstriction and atrial natriuretic peptide–mediated afferent arteriolar vasodilation strive to maintain glomerular filtration in the face of renal hypoperfusion.

Exercise

Exercise constrained by dyspnea or fatigue is a common heart failure symptom indicative of decreased oxygen delivery (central) or utilization (peripheral). Limiting central mechanisms include heart failure–associated reductions in ventilation, diffusion, chronotropic competence, and cardiac output. Also, β ₂ -adrenoceptor polymorphisms associated with impaired exercise performance have been identified. Proposed peripheral limiting mechanisms include sarcopenia, impaired mitochondrial respiration, and diminished oxygen transport due to augmented neurogenic vasoconstriction. In heart failure patients (but not in healthy age-matched controls) maximal oxygen uptake (peak VO ₂ ) during exercise correlates inversely to resting MSNA but not cardiac NE spillover.

In recent experiments, fibular MSNA was recorded continuously while subjects performed contralateral one-leg cycling without and against resistance. In contrast to healthy controls, whose burst incidence fell, exercise increased MSNA in those with HFrEF; overall there was a significant inverse relationship between the maximum MSNA elicited by one-leg exercise and subjects’ peak VO ₂ ( Fig. 13.5 ). Sympathetically mediated reductions in blood flow below levels required to meet local metabolic demands during exercise could attenuate endothelium-mediated vasodilatation and stimulate metaboreceptor afferents in skeletal muscle. Consequent reflexive increases in central sympathetic outflow could further diminish exercise capacity.