The autonomic nervous system has two primary branches: (1) the parasympathetic; and (2) the sympathetic nervous systems [21, 22]. The balance between parasympathetic and sympathetic nerve activity modulates cardiovascular homeostasis and the response to exercise [21, 23]. For example, at dynamic exercise onset, parasympathetic withdrawal causes a nearly instantaneous increase in heart rate. Conversely, increased sympathetic activation at exercise onset causes a gradual rise in heart rate. The dominant controller of heart rate thus shifts from parasympathetic actvity at rest (i.e., high activity) and during initial exercise onset (i.e., withdrawal) to sympathetic activity (“sympathoexcitation”) during sustained exercise[23, 24]

Exaggerated sympathoexcitation in HFrEF may play a large role in disease progression [25] and is a marker of prognosis and survival [20, 26]. At the onset of left ventricular systolic dysfunction, the balance between parasympathetic and sympathetic nerve activity shifts in response to: (a) an increase in left ventricular end-diastolic pressure; or (b) a decrease in stroke volume [20, 26, 27, 28]. This shift in balance favours sympathoexcitation to temporarily re-establish hemodynamic equilibrium [21]. As HFrEF advances, parasympathetic outflow progressively wanes while sympathetic outflow increases [29], independent of HFrEF etiology [30, 31]. Concurrently, renin-angiotensin-aldosterone system activity increases [30, 31]. These autonomic and neurohormonal responses may initially be “adaptive” to maintain cardiac function [21, 30], but may subsequently and progressively become maladaptive and worsen HFrEF severity.

Autonomic dysfunction in HFrEF combines cardiac, ventilatory, neurohormonal, and vascular factors [6, 28, 30, 32] that are also related to poor prognosis [33] and increased mortality [20, 34]. Cardiac factors may include decreased tonic and reflex vagal heart rate modulation, increased resting heart rate [34, 35], chronotropic incompetence [36], and decreased heart rate variability [9, 36, 37, 38]. Elevated resting minute ventilation [12] and ventilation-to-perfusion mismatch are also consistently found in HFrEF [6, 33, 39, 40]. Neurohormonal changes in HFrEF include increased venous plasma norepinephrine, cardiac and renal norepinephrine spillover, cardiac norepinephrine stores, and plasma brain natriuretic peptides [6, 41]. Increased renal sympathetic nerve activity reduces renal blood flow, which may in turn exacerbate the renin-angiotensin-aldosterone system [42]. Finally, increased resting muscle sympathetic nerve activity (MSNA) burst frequency in HFrEF [9, 10, 11, 12, 13, 14] elicits tonic vasoconstriction and increased peripheral resistance at rest. Exaggerated sympathoexcitation in HFrEF is most common with severe exercise intolerance [15, 19] or ischemic heart disease [9].

Contrary to initial conceptual models that emphasized a generalized, non-selective increase in sympathetic nerve activity to all vasculature in response to left ventricular systolic dysfunction [27], evidence suggests a selective alteration in autonomic cardiovascular changes in the early stages of HFrEF progression (e.g., increased cardiac norepinephrine spillover versus no change in total or renal norepinephrine spillover) [28]. Autonomic dysfunction in HFrEF may be examined as the exaggeration or blunting of cardiovascular regulating autonomic reflexes in response to exercise: the exercise pressor reflex, the arterial baroreflex, and the central and peripheral chemoreflexes [21, 22] (Fig. 1, Ref. [21]). The current review is focused on extra-cardiac reflexes, therefore, the cardiopulmonary baroreflex and cardiac chemoreflex are not discussed herein. We refer the reader to this review for further detail [43].

Fig. 1.

Integration of exercise-related, cardiovascular autonomic reflexes in HFrEF. Afferent feedback is collected and integrated at the medulla oblongata of the brain stem (blue hexagon), where from efferent signals are sent to various organs in response to the integration of the afferent and central information. Excitatory (+) and inhibitory (-) stimuli are shown for afferent feedback (black arrows), efferent sympathetic nerve activity (red arrows) and efferent parasympathetic nerve activity (green arrows). Exaggerated stimuli due to HFrEF are indicated as thick arrows, while blunted stimuli due to HFrEF are indicated as dashed thin arrows. *Note: the effects of sympathoexcitation on cardiac contractility in HFrEF are generally to increase contractility, but some research suggests stroke volume remains the same or decreases during the activation of the metaboreflex in HFrEF. **Note: the aortic bodies contain both baroreceptors and chemoreceptors, which emit contrasting afferent signals. The baroreflex stimulus is inhibitory and is blunted in HFrEF; the chemoreflex stimulus is excitatory and is exaggerated in HFrEF. Figure concept and representation of afferent and efferent pathways was inspired by figure 4 in reference [21].

The muscle mechanoreflex may independently contribute to HFrEF pathophysiology. Deformation of the skeletal muscle mechanoreceptors increases the muscle afferent signalling rate to the central nervous system [46]. Mechanoreceptors rapidly adjust their firing rate to changes in muscle tension consistency and respond earlier and faster than metaboreceptors [47]. The mechanoreflex elicits an increase in sympathoexcitation that increases arterial and renal vasoconstriction [58], and that debatably/negligibly increases ventilation [3, 4, 47]. Thus, muscle contraction indirectly increases perfusion pressure and O${}_2$ delivery to active tissues and prevents peripheral blood flow from surpassing cardiac output during exercise [46, 47]. Studies isolating the mechanoreflex from both the metaboreflex and central command are scarce [48] and may use passive muscle stretch, electrically stimulated involuntary muscle contractions, or low-intensity handgrip exercise (this includes central command).

Renal vasoconstriction during exercise is exaggerated during isolated mechanoreflex but not metaboreflex activation in HFrEF [58]. In contrast, the mechanoreflex has little effect on the exercise ventilatory response in both HFrEF and health, while the metaboreflex has a large regulatory role in ventilation modulation during exercise [3, 4]. In rodents with dilated cardiomyopathy (coronary ligation), the exercise pressor response and isolated mechanoreflex activity (i.e., mean arterial pressure and heart rate responses) progressively increased as left ventricular function decreased over a 10-week period [59]. Thus, an exaggerated mechanoreflex response in HFrEF may be systolic function dependent [59]. MSNA increased similarly during normal exercise in HFrEF compared to controls, both with dichloroacetate (which reduces lactic acid production) or saline (placebo) transfusion, but increased earlier during exercise in HFrEF compared to controls during placebo [50]. Further, mechanoreflex isolation via passive arm stretch caused an exaggerated early and sustained elevation of MSNA in HFrEF compared to no effect in controls [50]. Thus, there may be a prominent role for increased mechanosensitivity in the exaggerated sympathoexcitation during exercise in patients with HFrEF [50]. Additionally, the findings suggest muscle acidity does not affect mechanoreceptor sensitization in patients with HFrEF [50], which contrasts other reports [46]. In summary, the mechanoreflex may be exaggerated in HFrEF, such that renal vasoconstriction, mean arterial pressure, heart rate, and MSNA responses to exercise are elevated.

The muscle metaboreflex independently contributes to exaggerated sympathoexcitation in HFrEF [3, 4, 19, 47, 60]. The metaboreflex occurs when increased metabolite concentration (e.g., ATP [61], lactic acid, adenosine [10], phosphate, kinins, and cations [47, 62]) stimulates skeletal muscle receptors, whose afferent firing rate increases to yield hemodynamic, ventilatory, and sympathetic changes [21]; these changes include increases in MSNA, peripheral vasoconstriction [47], mean arterial pressure, pulmonary vascular resistance [63], and ventilation or hyperpnea [47]. During exercise, the metaboreflex increases cardiac output via stroke volume in health [64] and potentially heart rate [65] depending on the state of exercise, recovery, or ischemia [47]. However, the presence of HFrEF alters these effects.

The metaboreflex is consistently elevated in patients with HFrEF [21, 66] such that MSNA during ischemic and nonischemic handgrip exercise is greater and occurs earlier [19]; exercise ventilation is amplified due to an exaggerated metaboreflex response [3, 66, 67] in symptomatic but not in asymptomatic patients with HFrEF [68]; post-exercise heart rate recovery is slower [60]; and arterial pressure is elevated during post-exercise circulatory occlusion [60, 64, 67, 69, 70, 71]. While increased arterial pressure is primarily due to increased cardiac output in health, metaboreflex activation in HFrEF elicits negligible cardiac output changes and thus elevates mean arterial pressure via peripheral vasoconstriction and elevated systemic vascular resistance [63, 64, 69, 70, 71]. A dose-response relationship for this phenomenon was recently reported across three metaboreflex exercise intensities [72]. Cardiac output was limited during metaboreflex activation because of increased coronary vasoconstriction in dogs with HFrEF, but this has yet to be confirmed in humans with HFrEF [73]. Thus, the metaboreflex mechanisms in HFrEF may be related to the impaired ability to increase stroke volume during exercise [64, 70, 71].

In HFrEF, metaboreflex sensitivity and activity are negatively related to peak VO${}_2$ [19, 67] and heart rate variability (indirect measure of cardiac autonomic regulation and associated with higher mortality [67, 74]), and are positively related to the exercise ventilatory response [51, 67], circulating catecholamine levels [67], and mortality risk score [68]. Further, metaboreflex activity is not related to ventricular function [67], and yet it increases as HFrEF severity worsens [51]. Metaboreflex exaggeration in HFrEF may be associated with increased peripheral, but not central, chemoreflex sensitivity and decreased arterial baroreflex regulation of heart rate, independent of left ventricular EF [67].

Some research suggests that HFrEF, perhaps only in its more severe form, attenuates [46] or has no effect on the metaboreflex [50]. Patients with mild HFrEF and controls have elevated MSNA and blood pressure during metaboreflex activation but not in those with severe HFrEF [75]. Further, one HFrEF group observed a decrease in MSNA from exercise to post-exercise ischemia while controls observed an increase [11]. The same group later published directly contrasting results in addition to earlier fatigue and greater accumulation of muscle metabolites in HFrEF, thus supporting the overall metaboreflex exaggeration in HFrEF [76]. Conflicting results on metaboreflex activity in HFrEF may result from different measured physiological indices (e.g., MSNA vs. ventilation vs. blood pressure) [77], exercise modes (e.g., rhythmic or isometric), muscle group (e.g., forearm or knee extensor), and/or HFrEF severity (e.g., muscle quality, muscle afferent sensitization or desensitization). For more on this, we refer the reader to the Wang et al. [77] and the give and take discussion between the blunted and exaggerated hypotheses. Taken together, we know muscle metaboreflex dysfunction in HFrEF alters exercise and post-exercise MSNA (mixed views on blunted vs. exaggerated responses [47]), and increases exercise and post-exercise blood pressure, vasoconstriction, and ventilation. Thus, dysfunctional metaboreflex may play a critical role in manifesting HFrEF exaggerated sympathoexcitation, early fatigue, and exercise intolerance.

The central chemoreceptors are located adjacent to the respiratory centres in the medulla oblongata and detect blood CO${}_2$ and pH levels. When blood CO${}_2$ levels rise (“hypercapnia”), there is more CO${}_2$ in the interstitial space surrounding the central chemoreceptors, triggering afferent signals to the respiratory centres in the medulla oblongata. The respiratory centres respond by raising ventilation. Studies targeting the central chemoreflex activate the CO${}_2$-sensitive central chemoreceptors while simultaneously deactivating the O${}_2$-sensitive peripheral chemoreceptors by breathing hyperoxic hypercapnic gas (93% O${}_2$ and 7% CO${}_2$ [83]). Patients with HFrEF have greater ventilation [13], heart rate [13], and MSNA [12, 13] and decreased forearm blood flow [12] during resting hyperoxic hypercapnia compared to controls. Further, symptomatic patients have greater MSNA and ventilation and attenuated sympathoinhibition during hyperoxic hypercapnia compared to asymptomatic HFrEF patients [106]. These findings are clinically relevant as blood CO${}_2$ levels increase during exercise and/or sleep apnea, therefore, in patients with HFrEF, elevated blood CO${}_2$ levels exaggerate the central chemoreflex and exacerbate the already elevated sympathoexcitation. Clinically, hyperventilation is more prognostic than peak VO${}_2$ in HFrEF [83], although both are meaningful endpoints. Abnormal exercise hyperventilation in HFrEF is often due to high ventilation-to-perfusion mismatch and increased VCO${}_2$ due to greater bicarbonate buffering of lactic acid [107]. Even so, there is more data on the peripheral rather than central chemoreflex during exercise in HFrEF.

Exaggerated peripheral chemoreflex activity is common in HFrEF and may contribute to exaggerated sympathoexcitation, hyperventilation, breathing instability (e.g., Cheyne-Stokes) [102, 108], and autonomic dysfunction in HFrEF [30]. The peripheral chemoreceptors (“glomus cells”) in the carotid bodies and aortic arch detect blood O${}_2$ levels. When blood O${}_2$ levels fall, the chemoreceptors depolarize the glossopharyngeal and vagus nerves, signaling the medulla oblongata to increase ventilation [102], increase heart rate [103], and increase vasoconstriction to maintain perfusion pressure [12, 102]. The peripheral chemoreflex is investigated using steady-state (10% O${}_2$ and 90% N${}_2$ [12]) or transient (100% N${}_2$ [109]) hypoxia. Titrating CO${}_2$ uptake maintains arterial isocapnia to minimize central chemoreflex activity [12].

Hypoxemia increases ventilation at rest and during exercise, and decreases the peak work rate but not peak VO${}_2$ in HFrEF [39], the latter suggesting greater work of breathing with hypoxia. Patients with HFrEF had greater MSNA and reduced forearm blood flow compared to controls during resting isocapnic hypoxia [12]. Carotid chemoreceptor inhibition via dopamine caused resting hindlimb vasodilation in HFrEF but not healthy dogs, while both groups experienced similar exercise vasodilation [110]. By adding alpha-adrenergic blockade with phentolamine, carotid chemoreceptors accounted for roughly 1/3rd of the sympathetic tone during exercise in healthy dogs [110] while this additional blockade abolished exercise vasodilation in HFrEF dogs [110]. Notably, carotid body denervation in two healthy dogs abolished the exercise vasodilation [110]. Thus, peripheral chemoreflex sensitivity is exaggerated in HFrEF [102]. Indeed, animal models indicate that the transformation from normal, tonic chemoreceptor activity to HFrEF-characteristic chemoreceptor hyperactivity may occur in a matter of weeks [102].

The carotid chemoreflex interacts with the muscle metaboreflex to regulation minute ventilation in a breathing frequency dependent manner in patients with HFrEF, potentially contributing the exercise hyperventilation and exercise intolerance in HFrEF [111]. In other words, the activation of the muscle metaboreflex worsens the effects of the already hyperactive carotid chemoreflex in patients with HFrEF; as exercise stimulates both reflexes, this interaction may be a key player in functionally limiting exercise tolerance, which in turn is related to quality of life and mortality in HFrEF. The interaction between these situationally distant reflexes poses a therapeutic problem pharmacologically, however, exercise therapy (see section 1.6) may be used to address both reflexes individually and thus improve the nature of their interaction in patients with HFrEF.

Carotid body ablation, resection, or denervation have demonstrated improved autonomic function (reduced renal sympathetic nerve activity or MSNA, increased renal blood flow, increased heart rate variability, reduced peripheral chemosensitivity, increased baroreflex sensitivity) in animal HFrEF models [112, 113, 114]. A 2017 first-in-man study demonstrated that carotid body resection (4 patients unilateral resection, 6 patients bilateral resection) in HFrEF decreased both MSNA and peripheral chemosensitivity and improved exercise tolerance 1-month post-resection [115]. However, safety concerns arose in those with bilateral resection who trended lower of oxygen saturation at night [115]; carotid bodies stimulate protective reactions to hypoxia, therefore, further investigation balancing patient safety and clinical benefit is required prior to wider usage of carotid body resection or alteration. Although still debated, hemodynamic, neurohumoral, and local metabolic mechanisms underlying HFrEF pathophysiology may collectively progress peripheral chemoreceptor sensitization [102]. Despite the high O${}_2$ supply to peripheral chemoreceptors in health, it is speculated that relatively reduced O${}_2$ supply in HFrEF may play a role in peripheral chemoreceptor sensitization [102]. As previously mentioned, hyperventilation, sleep apneas, and Cheyne-Stokes respiration are common in patients with HFrEF and are secondary to peripheral chemoreflex exaggeration, not vice versa [102]. However, these ventilatory abnormalities may also exacerbate the mechanisms of chemoreflex exaggeration [102]. Oxidative stress may stimulate carotid chemoreceptor sensitization in HFrEF [32, 102]. Chronic exposure to intermittent hypoxemia, as is common in HFrEF, sensitizes the carotid chemoreceptors to be more “excitable” to acute hypoxemia in HFrEF rats, likely due, in part, to greater carotid body oxidative stress [116]. Angiotensin II, a pro-oxidant, is chronically elevated in ~45–50% of patients with severe HFrEF (NYHA class III–IV) despite ACE inhibitor treatment [117, 118] and is associated with poor prognosis in HFrEF [117]. The NADPH-angiotensin II-superoxide pathway may increase carotid body chemoreceptor sensitivity to hypoxemia in humans with HFrEF [119]. Finally, comorbidities common in HFrEF are also related to increased oxidative stress and may exacerbate carotid body sensitization [32]. Therefore, the role of oxidative stress remains an essential component of HFrEF pathophysiology. In summary, the peripheral chemoreflex is exaggerated in patients with HFrEF and contributes to abnormal ventilation.

It is imperative to consider that improved exercise capacity, cardiac function, and/or skeletal muscle function does not equal improved autonomic function. Similar to how exercise regime (moderate intensity continuous training, high intensity interval training, resistance training, etc.) yields different health and clinical benefits improve autonomic function, although the majority include some form of aerobic training. Low intensity cardiac rehabilitation in patients with HFrEF [143], whether or not the exercise regime improves autonomic function in HFrEF may depend greatly on the mode, duration, and intensity of exercise [30]. Nonetheless, it appears that a wide variety of regimes restored autonomic tone and reactivity to sympathetic and parasympathetic stimuli in patients with HFrEF [137]. An aerobic, resistance-band training study targeting each muscle group one at a time lowered circulating catecholamine levels at rest and during exercise in HFrEF patients [144]. Combined interval and resistance training had statistically similar effects on respiratory drive compared to interval training alone in patients with HFrEF, but data indicated that combined training may have improved resting and exercise breathing pattern indices and P${}_{ET}$CO${}_2$ more than intervals alone [145]. Inspiratory muscle training has produced benefits to exercise tolerance and potentially autonomic function in patients with HFrEF [146]. Clearly, exercise training program parameters (frequency, intensity, type, and time) are highly varied among autonomic studies in patients with HFrEF. Therefore, studies explicitly evaluating autonomic measures and mechanisms are required in and above clinical outcomes.

Exercise training alters both the muscle metaboreflex and the muscle mechanoreflex in patients with HFrEF [147, 148]. Six weeks of forearm training reduced the systolic and diastolic pressure, minute ventilation, respiration rate, and leg vascular resistance responses and increased the leg blood flow response to post-handgrip exercise circulatory occlusion in patients with HFrEF more so compared to healthy controls [148]. Further, these changes were accompanied by a decrease in sympathetic activation but without change in parasympathetic activation as measured by power spectral analysis of blood pressure low frequency components [148]. Antunes-Correa et al. [147] reported lower resting MSNA following exercise training (4-months combined aerobic, resistance, and flexibility training) compared to an untrained control HFrEF group; interestingly, however, the nature of the metaboreflex improvement further complicates the known literature surrounding muscle metaboreflex dysfunction in patients with HFrEF. During muscle metaboreflex activation via post-exercise circulatory occlusion, exercise trained patients with HFrEF reported heightening MSNA as well as increased transient receptor potential vanilloid type-1 (TRPV1) receptors and cannabinoid receptor type-1 (CB1) in the vastus lateralis (trained muscle) whereas the untrained group saw no change in MSNA or the expression of either receptor compared to baseline testing [147]. This latter finding suggests that the blunted muscle metaboreflex response, found in some HFrEF studies [11, 75, 149] but not all [21, 47, 51, 66, 148], and confirmed at baseline testing in this study, is augmented with exercise training, thereby returning muscle metaboreflex dysfunction closer to normal function. It may be that HFrEF severity dictates the nature of muscle metaboreflex dysfunction [75], such that mild cases (e.g., NYHA class I—II) exhibit exaggerated responses that deteriorate as disease severity worsens, resulting in a blunted muscle metaboreflex response in more severe HFrEF conditions (e.g., NYHA class III–IV). In addition to muscle metaboreflex isolation, authors isolated the mechanoreflex with passive exercise and found decreased MSNA and expression of cyclooxygenase-2 (COX-2), prostaglandin-E2 receptor-4, and thromboxane-A2 receptors [147], each of which is involved in the COX pathway, a hypothesized mechanism behind mechanoreflex dysfunction in HFrEF. Recently, 6-months of exercise training cardiac rehabilitation in patients with HFrEF not only improved peak VO${}_2$, resting and low-intensity exercising MSNA, but also attenuated the exaggerated MSNA during muscle metaboreflex activation compared to pre-training levels, although remained greater than healthy controls both pre- and post-training [150]. Aside from these high-impact data, little research has focused on the effect of exercise training on the exercise pressor reflex, the isolated muscle metaboreflex, or the isolated muscle mechanoreflex in patients with HFrEF. However, inferences may be made from animal HFrEF studies.

Two studies by Wang et al. [151, 152] demonstrated that exercise training targets both the muscle metaboreflex and mechanoreflex in rat HFrEF models. Exercise training normalized pre-exercise blunted muscle metaboreflex (isolated via capsaicin injection) and decreased the pre-exercise exaggeration of the muscle mechanoreflex (isolated via contraction and passive stretch) in rats with HFrEF, with no effects seen in sham rats post-training [152]. Mechanistically, exercise training improved the sensitization of both group III and IV afferents in the HFrEF rat models, corresponding to improved function of certain muscle metaboreflex (TRPV1) and mechanoreflex (P2X) receptors [151]. A review by the same group discussed the discrepancy between the observations of both blunted and exaggerated muscle metaboreflex responses in HFrEF, as well as potential mechanisms underlying the improved exercise pressor reflex response after exercise training, to a greater detail than may be covered herein [77].

The effect of exercise training on baroreflex function in patients with HFrEF is, surprisingly, vastly understudied in humans. To the knowledge of the authors, the only known study in human patients with HFrEF showing improved baroreflex sensitivity post-exercise training lacked a control group [153]. Fortunately, 6-months of combined aerobic and circuit training significantly improved baroreflex sensitivity with concomitant improvements in heart rate variability, cardiac presynaptic sympathetic innervation, increases in myocardial blood flow where it was reduced pre-training, however, without a statistically significant increase in peak VO${}_2$, in 12 NYHA class II—III HFrEF patients [153]. The very small sample and lack of a patient control group still enforce the need for updated evidence in baroreflex function post-exercise training in patients with HFrEF. Similarly, simulating exercise training with peroneal and tibial nerve electrical stimulation (transcutaneous electrical nerve stimulation; TENS) in patients with HFrEF effectively improved baroreflex sensitivity [96, 154]. In contrast to the TENS groups, however, brief (2-day) exercise “training” had no effect on baroreflex sensitivity [96] and two weeks (4–6 sessions, 75-min each) of exercise training still failed to significantly improve baroreflex sensitivity in patients with HFrEF [154].

Despite the surprising lack of controlled, exercise training trials in patients with HFrEF on baroreflex sensitivity, evidence strongly suggests that exercise therapy improves baroreflex function in patients on the pre-HFrEF spectrum (e.g., hypertension, coronary artery disease, post-myocardial infarction, etc. [155, 156]) as well as in animal HFrEF models [128, 129, 157, 158, 159], and thus strongly warrants clinical trials investigating the efficacy of improved baroreflex function and corresponding health benefits from exercise. However, given the lack of patient evidence, this section of the review will focus on animal based HFrEF studies.

Exercise training improved arterial baroreflex control of both renal sympathetic nerve activity and heart rate in HFrEF rabbits [158] and rats [160]. The latter finding was accompanied by improved aortic depressor nerve activity sensitivity following exercise training in rats with HFrEF [160], suggesting a role not only of carotid baroreceptors but of aortic baroreceptors in modulating heart rate and sympathetic outflow in HFrEF. Indeed, work by Liu et al. [159] demonstrated that atropine and metoprolol infusion modulated baroreflex control of heart rate through increased vagal tone, rather than reduced sympathetic outflow, following exercise training in HFrEF.

Exercise training normalized sympathetic outflow (resting renal sympathetic nerve activity) and arterial baroreflex function (attenuated maximum gain of heart rate and renal sympathetic nerve activity curves) in rabbits with HFrEF [128, 157]. Further, angiotensin II type 1 receptor blockade in non-exercise trained HFrEF rabbits raised baroreceptor sensitivity approximating that of those exercise-trained rabbits with HFrEF; exercise in HFrEF rabbits also lowered circulating angiotensin II levels [157]. Mousa et al. [158] attributed their observed reduction in renal sympathetic nerve activity and elevation in baroreflex sensitivity in rabbits with HFrEF to concurrent reduction in circulating angiotensin II and angiotensin receptors in the central nervous system that occurs following exercise training. Thus, angiotensin II and the RAAS pathway is mechanistically tied to the dysfunctional pathophysiology of the arterial baroreflex in HFrEF. In rats, only endurance training improved cardiac autonomic control (increased baroreflex sensitivity) compared to moderate interval training, despite improved cardiac function and exercise capacity in both programs [129].

Patients with HFrEF following carotid body resection exhibited significantly improved exercise time compared to pre-resection exercise time, reduced MSNA, and decreased peripheral chemosensitivity [115]. Although these are encouraging findings in human patients, this approach has been contested, for many reasons, in favour of exercise training [161]. Exercise training poses a less-invasive, more cost-effective, and in most cases, simpler alternative that may produce similar or better results, as well as more global physiological and health benefits (endothelial function, peripheral blood flow distribution, skeletal muscle factors, and metabolic factors associated with reduced mortality).

Animal models have confirmed the efficacy of exercise training on attenuating exaggerated chemoreceptor sensitivity in HFrEF. Exaggerated peripheral chemosensitivity was attenuated by exercise training in rabbits with HFrEF [162]. Renal blood flow is classically reduced in the HFrEF condition due to exaggerated renal sympathetic nerve activity and consequent renal vasoconstriction, mediated in part by heightened peripheral chemoreceptor sensitivity. Exercise training attenuated peripheral chemoreflex-mediated tonic reductions in renal blood flow in rabbits with HFrEF, as assessed by compared carotid-body chemoreceptor denervated animals with carotid-body intact animals; further, exercise training also attenuated hypoxia induced reductions in renal blood flow in the same HFrEF sample [163]. Along with improved exercise capacity, rabbits with pacing-induced HFrEF returned to pre-pacing levels of renal blood flow and vascular conductance during acute hypoxia after exercise training [127]. Thus, animal models have demonstrated that exercise training attenuates exaggerated peripheral chemoreceptor sensitivity found in HFrEF and improves physiological functions (increased peripheral blood flow) and autonomic signalling (reduced renal sympathetic nerve activity, increase baroreflex sensitivity). Unfortunately, in human patients with HFrEF, data supporting direct involvement and improvement of peripheral chemoreceptor function is lacking. Both interval training and combined interval and resistance training improved respiratory drive in patients with HFrEF [145]. Coats et al. [134] demonstrated reduced minute ventilation and the ventilation-carbon dioxide production slope post-exercise training, along with improved heart rate variability and reduced catecholamine levels, indicative of a shift in sympathovagal from sympathetic toward greater vagal tone, in patients with HFrEF.