Carotid body modulation in systolic heart failure from the clinical perspective

Abstract

Augmented sensitivity of peripheral chemoreceptors (PChS) is a common finding in systolic heart failure (HF). It is related to lower left ventricle systolic function, higher plasma concentrations of natriuretic peptides, worse exercise tolerance and greater prevalence of atrial fibrillation compared to patients with normal PChS. The magnitude of ventilatory response to the activation of peripheral chemoreceptors is proportional to the level of heart rate (tachycardia) and blood pressure (hypertension) responses. All these responses can be measured non‐invasively in a safe and reproducible fashion using different methods employing either hypoxia or hypercapnia. Current interventions aimed at modulation of peripheral chemoreceptors in HF are focused on carotid bodies (CBs). There is a clear link between afferent signalling from CBs and sympathetic overactivity, which remains the priority target of modern HF treatment. However, CB modulation therapies may face several potential obstacles: (1) As evidenced by HF trials, an excessive inhibition of sympathetic system may be harmful. (2) Proximity of critical anatomical structures (important vessels and nerves) makes surgical and transcutaneous interventions on CB technically demanding. (3) Co‐existence of atherosclerosis in the area of carotid artery bifurcation increases the risk of central embolic events related to CB modulation. (4) The relative contribution of CBs vs. aortic bodies to sympathetic activation in HF patients is unclear. (5) Choosing optimal candidates for CB modulation from the population of HF patients may be problematic. (6) There is a risk of nocturnal hypoxia following CB ablation – mostly after bilateral procedures and in patients with concomitant obstructive sleep apnoea.

Abbreviations

AB

aortic body

CB

carotid body

HF

heart failure

HR

heart rate

HVR

hypoxic ventilatory response

$P_{\mathrm{aC}{\mathrm{O}}_2}$

partial pressure of carbon dioxide in arterial blood

$P_{\mathrm{a}{\mathrm{O}}_2}$

partial pressure of oxygen in arterial blood

PChR

peripheral chemoreceptors

PChS

peripheral chemosensitivity

SBP

systolic blood pressure

$S_{\mathrm{p}{\mathrm{O}}_2}$

peripheral capillary oxygen saturation

Introduction

An augmented reflex response from peripheral chemoreceptors (PChRs) is a common feature of patients with systolic heart failure (HF). It has been showed that afferent signalling from PChRs is related to excessive activation of the sympathetic nervous system (Despas et al. 2009, 2012; Franchitto et al. 2010), which is perceived as a priority target for modern HF treatment (Floras & Ponikowski, 2015). Moreover, both enhanced peripheral chemosensitivity (PChS) (Ponikowski et al. 2001) and increased sympathetic tone (Cohn et al. 1984) are well recognized markers of poor long‐term prognosis in the HF population. Thus, a concept of PChR modulation for the treatment of systolic HF has been established (Niewiński et al. 2013; Paton et al. 2013).

Transient inhibition of PChRs with low‐dose dopamine (by activation of D₂ receptors on type I glomus cells) or hyperoxia (by interaction with K⁺ channels) was found to acutely improve exercise capacity in healthy volunteers (Janssen et al. 2009) and in patients with systolic HF (Chua et al. 1996). A decrease in noradrenaline levels and improvements in haemodynamic indices following bromocriptine administration (a potent D₂ receptor agonist) further support the idea of PChR modulation in stable HF (Francis et al. 1983). Subsequently, a case report of unilateral CB resection demonstrated augmented exercise parameters (higher peak oxygen consumption and lower slope relating ventilation to carbon dioxide production) in a single patient with advanced systolic HF (Niewiński et al. 2013). This effect might be related to the attenuation of sympathetic constraint towards skeletal muscle arterioles as described in an animal model (Alves et al. 2007). On the other hand, it could be hypothesized that exercise improvements are related to the subjective perception of diminished dyspnoea, which has been shown following CB resection for respiratory disorders (Vermeire et al. 1987) and after oxygen administration in HF (Moore et al. 1992).

In humans there are two peripheral chemosensory areas, i.e. carotid bodies (CBs) and aortic bodies (ABs). However, only CBs may be viewed as a reliable target for invasive intervention due to their relatively homogenous anatomical distribution (bilaterally in the area of common carotid artery bifurcation) (Heath et al. 1970; Kumar & Prabhakar, 2012).

Prevalence of augmented peripheral chemosensitivity

In clinical studies, PChS is defined as augmented if its numerical value exceeds mean + two standard deviations based on data obtained from an age‐matched and gender‐matched healthy population. This equates to cut‐offs between 0.68 and 0.77 l min⁻¹ per peripheral capillary oxygen saturation ($S_{\mathrm{p}{\mathrm{O}}_2}$) (Chua et al. 1997; Giannoni et al. 2008; Niewinski et al. 2013) for the most commonly employed assessment methods. When using the above definition, the prevalence of augmented PChS in HF population is estimated at 40–44%. Importantly, the prevalence of elevated PChS has not changed significantly over the last two decades, despite significant advances in the pharmacotherapy of HF (see below) and despite introduction of effective coronary and electrophysiological interventions (Chua et al. 1997; Giannoni et al. 2008; Niewinski et al. 2013). This underscores the futility of currently available treatment modalities in correcting autonomic imbalances seen in systolic HF.

Influence of heart failure pharmacotherapy on peripheral chemosensitivity

Drugs currently recommended for the treatment of systolic HF (McMurray et al. 2012) may potentially influence PChS. In the study by Janssen et al. digoxin increased peripheral chemosensitivity (Janssen et al. 2010). Moreover, animal studies suggest that enhanced CB chemoreceptor activity might be mediated by angiotensin II and AT‐1 receptors (Li et al. 2006), which are targets of angiotensin converting enzyme inhibitors (ACE‐I) and angiotensin receptor blockers (ARBs) commonly used in HF. In contrast, beta‐blockers (also widely used in HF) do not influence peripheral chemosensitivity (Beloka et al. 2008).

Clinical importance of augmented peripheral chemosensitivity (Fig. 1)

Figure 1.

Clinical phenomena related to augmented peripheral chemosensitivity in heart failure

Patients with high PChS are characterized by low sensitivity of arterial baroreceptors, which might be the consequence of close anatomical proximity of structures within the central nervous system that are involved in both reflex arcs, such as nucleus tractus solitarii, hypothalamic defence area, reticular formation and ventrolateral medulla oblongata. This relation is reciprocal, as baroreceptor stimulation inhibits chemoreceptor activation and chemoreceptor excitement diminishes baroreflex‐mediated inhibition of the sympathetic system (Wennergren et al. 1976; Marshall, 1981; Ponikowski et al. 1997). In HF syndrome, low baroreflex sensitivity constitutes a well‐established predictor of poor prognosis and haemodynamic instability (Osterziel et al. 1995; Mortara et al. 1997).

Furthermore, HF patients with augmented level of PChS are prone to ventricular and supraventricular arrhythmias including atrial fibrillation (Chua et al. 1997; Giannoni et al. 2008). This may be caused by co‐existence of an increased sympathetic tone known to promote cardiac dysrhythmias in the HF population (Meredith et al. 1991). Alternatively, it may be related to more advanced morphological changes of the myocardium increasing the probability of proarrhythmic substrate formation (Benes et al. 2011).

Peripheral chemosensitivity is also augmented in HF patients with central sleep apnoea (Kara et al. 2003), where it contributes to cyclic fluctuations in carbon dioxide concentration resulting in Cheyne‐Stokes ventilation (Hanly et al. 1993). Moreover, exaggerated PChS explains recurrent surges in sympathetic tone with concurrent tachycardia and hypertension. These lead to cyclic rises in afterload and myocardial oxygen demand, resulting in repetitive cardiac ischaemia (Kaye et al. 1995).

Assessment of peripheral chemosensitivity

There are several methods for reliable and reproducible assessment of PChS. To elicit an acute response from PChR (i.e. reaction to sudden stimulation) various research groups employ either hypoxia or hypercapnia in poikilocapnic, isocapnic or isooxic conditions (Niewinski, 2014). Generally, the level of PChS is calculated based on the magnitude of change in minute ventilation in proportion to the decrease in $S_{\mathrm{p}{\mathrm{O}}_2}$ or increase in $P_{\mathrm{ET},\mathrm{C}{\mathrm{O}}_2}$ (end‐tidal $P_{\mathrm{C}{\mathrm{O}}_2}$). While technical aspects of these tests may vary, the interpretation remains similar. Importantly only two methods (i.e. the acute intermittent hypoxic test with nitrogen and the isocapnic hypoxic method using rebreathing) have been found to have significant prognostic impact (Ponikowski et al. 2001; Giannoni et al. 2009) and thus can be recommended for clinical use in HF patients.

To assess the level of tonic (normoxic) activity of PChRs, either low‐dose dopamine (Niewinski et al. 2014 b) or a high concentration of oxygen (Sinski et al. 2012) may be used. Dopamine applied in vitro in a single injection to cat CB results in transient inhibition of chemosensory activity, whereas repeated administrations produce either a biphasic effect, no change or long‐lasting excitation (Zapata, 1975). It is now understood that dopamine in low dose inhibits the release of neurotransmitters from chemosensory cells by interacting with presynaptic D₂ receptors (Gonzalez et al. 1995), thus causing a decrease in minute ventilation which is proportional to the level of tonic PChS (Zapata & Zuazo, 1980; van de Borne et al. 1998). In contrast, large doses of dopamine have been reported to produce excitatory action which in a study by Herman et al. was antagonized by serotonin receptor type 3 (5‐HT₃) inhibitor but not by dopamine receptor antagonist (Herman et al. 2003). Brief administration of pure oxygen has been also proposed as a tool for the assessment of the tonic influence that PChRs exert upon ventilation (Dejours, 1962). Hyperoxia, while being convenient to employ, may with prolonged administration increase systemic vascular resistance and stimulate ventilation (Becker et al. 1996; Haque et al. 1996), making lengthy experimental protocols less reliable. At the moment it is unclear whether acute and tonic responses from PChR correlate (and therefore may be used interchangeably in HF patients) or constitute completely different entities. Thus, it is unknown which testing modality (measuring acute responses vs. tonic drive assessment) is best suited for the potential candidates for CB modulation.

The haemodynamic response (tachycardia and hypertension) to activation of PChRs is well described in the HF population, where it has been found to correlate with the magnitude of ventilatory reaction to hypoxia (Niewinski et al. 2013). This further confirms the co‐existence of enhanced sympathetic tone in HF patients with elevated PChS.

Clinical predictors of high peripheral chemosensitivity

Heart failure patients with augmented PChS can be reasonably well characterized from the clinical point of view. In terms of cardiac function such individuals present with lower left ventricle ejection fraction, greater diameter of right ventricle and greater diameter of left atrium compared to HF patients with normal PChS (Chua et al. 1997; Niewinski et al. 2013). Concomitantly, patients with increased PChS are characterized by lower peak oxygen consumption, greater slope relating ventilation to carbon dioxide production and higher New York Heart Association functional classification (measures of worse exercise capacity; Chua et al. 1997; Giannoni et al. 2008). Finally, patients with high PChS have higher levels of natriuretic peptides and greater prevalence of atrial fibrillation (Giannoni et al. 2008; Niewinski et al. 2013). All the above factors strongly point at the close association between HF advancement and the level of PChS, suggesting that high activity of PChRs is rather a result than a cause of HF syndrome (Ding et al. 2011). However, recent animal experiments suggested that CB hyperactivity not only has a compensatory function in HF (to maintain adequate ventilation and tissue perfusion in response to low peripheral blood flow), but also plays a pivotal role in disease progression (Del Rio et al. 2013; Marcus et al. 2014).

Choosing optimal candidates for CB modulation

There are several problems related to the optimal selection of the candidates for CB modulation in systolic HF. While augmented PChS ought to be among the major entry criteria for CB intervention, it is currently unknown: (1) what the cut‐off value for PChS should be, (2) which method should be employed for the assessment of PChS and (3) whether acute response and/or tonic activity ought to be tested. The list of potential contraindications to CB intervention is also unspecified with regard to the degree of carotid atherosclerosis or the severity and phenotype (obstructive vs. central) of sleep disordered breathing. Finally, it is unclear which patients on the spectrum of HF severity might benefit most from CB modulation. Would it be patients with less pronounced symptoms in which there is a degree of disease reversibility or should it be reserved for patients with advanced or even end‐stage illness – possibly as a form of palliative therapy when other measures fail? At the moment these questions remain without a definitive answer.

Potential problems related to carotid body modulation

While CB ablation may constitute a promising avenue of modern HF treatment, it has several potential limitations which are discussed below.

Excessive inhibition of sympathetic system

Results of the MOXCON trial utilizing moxonidine (a centrally acting imidazoline agonist inhibiting the sympathetic nervous system) suggested that excessive generalized sympathetic suppression may be harmful in the HF population (Cohn et al. 2003). The study was terminated prematurely due to excess mortality rate in the arm receiving moxonidine. There are several potential explanations for the results of the MOXCON trial (Floras, 2002). First, the baseline levels of sympathetic activation might have been too low for the study participants to benefit from potent sympathetic inhibition. Secondly, markedly reduced sympathetic tone could have been insufficient to support cardiac output and/or peripheral resistance, leading to progressive pump failure. Interestingly, in some patients receiving moxonidine in the MOXSE trial (a smaller study preceding the MOXCON trial) plasma noradrenaline levels were as low as the 25th percentile of the values reported for healthy controls (Swedberg et al. 2002). Thirdly, the rebound phenomenon was possibly in play, leading to uncontrollable rises in sympathetic activity causing increases in heart rate, blood pressure and aggravating ventricular ectopy. These observations suggest that CB modulation should be reserved for HF patients with not only augmented PChS but also confirmed hyperactivation of the sympathetic nervous system (e.g. measured with microneurography).

Anatomical considerations

Carotid bodies are located bilaterally in the area of common carotid artery bifurcation, usually closer to the external carotid artery. These are small but macroscopically discernible structures with cross‐sectional area of approximately 3 mm² (Nair et al. 2013; see Fig. 2). Apart from carotid arteries directly supplying the central nervous system, a number of important cranial nerves lie in close proximity to CBs. These include: vagus nerve, glossopharyngeal nerve, hypoglossal nerve, recurrent laryngeal nerve and branches of facial nerve. Intraoperative damage to these structures might be life threatening or cause severe disability, making a surgical approach rather challenging – especially in patients with short necks and/or obesity. However, among the vast number (>15000) of CB resections performed historically (for palliative treatment of respiratory disorders) only 13 cases of perioperative death were described within 48 h following the surgery. Nerve injury (mostly transient hypoglossal nerve paresis) was reported in up to 10% of cases and vessel injury in up to 9.5% of cases (Paton et al. 2013). The possibility of surgically inflicted complications led to the concept of percutaneous (transvascular) ablation of CBs. A specially designed radiofrequency catheter advanced through the common carotid artery could possibly be used to reach and destroy CBs. This approach, however, is also not without limitations. These include: local complications related to the vascular access, the need for parenteral anticoagulation and the risk of central nervous system embolization.

Figure 2. Macroscopic and microscopic structure of carotid body.

A, carotid body pinched by operative forceps (arrow) in the area of bifurcation of common carotid artery. B, microscopic examination of resected carotid body (H + E stain, magnification ×20). One of the multiple glomoids containing chemosensitive cells is indicated by the arrow.

Since the carotid sinus containing arterial baroreceptors is located adjacent to the carotid body, the risk of irreversible damage (or activation due to local scarring) of these structures must be also taken into consideration. Iatrogenic baroreflex dysfunction might result in unexpected and prominent fluctuations in blood pressure (Heusser et al. 2005). This is of particular importance for a radiofrequency approach where no direct visualisation is available during procedure. However, the surgical method proposed by Winter was specifically designed to avoid such complication (Winter, 1972).

Atherosclerosis in the region of carotid body

Patients with systolic HF are at increased risk of developing carotid atherosclerosis (Effoe et al. 2014), particularly when the aetiology of HF is ischaemic. Due to sheer‐stress distribution, the most common location of carotid atherosclerosis is the area of carotid bifurcation (Zarins et al. 1983) where CBs can be also found. The presence of atheromatous plaque is an obvious risk factor for peri‐procedural stroke which pertains to both surgical manipulation and transarterial radiofrequency ablation. Rates of central nervous system embolization reported in the literature (up to 1.8% of cases) might be underestimated, being based on historical data on CB resections performed in a specific population of patients with various respiratory disorders but not with systolic HF (Paton et al. 2013). Perhaps percutaneous techniques employing venous access to the area of carotid bifurcation may reduce the risk associated with the dislodgement of atherosclerotic material.

Role of aortic bodies

Peripheral chemoreceptors comprise CBs and ABs. All currently available methods for the assessment of PChS take into account summarized responses from both chemosensory areas. Therefore, it is unknown what is the relative contribution of CBs vs. ABs to the magnitude of PChS and to the overall sympathetic activation in HF. Nonetheless, it can be speculated, based on the observations made in HF patients following bilateral CB resection, that the ventilatory response is mediated solely by CBs, the blood pressure response mostly by CBs, and the heart rate response mainly by ABs (Niewinski et al. 2014 a; see Fig. 3). Recently, adenosine injected through a catheter placed in the common carotid artery (close to the bifurcation) has been proposed as a way of selective activation of CBs in humans. A similar (but less selective) approach was taken by Watt et al., who infused adenosine at various sites in the thoracic aorta (Watt et al. 1987). When adenosine was administered proximally to the origin of the carotid arteries an increase in minute ventilation was noted. Of interest, intra‐aortic adenosine infusion also resulted in heart rate augmentation irrespective of the site of infusion. This again speaks for the role of ABs in heart rate response to PChR activation. Such experiments are possible due to the extremely short half‐time of adenosine and its inability to cross the blood–brain barrier. The latter was elegantly showed in an animal study performed by Berne et al. where no significant radioactivity was detected in the brain tissue and cerebrospinal fluid following administration of ¹⁴C‐labelled adenosine into internal carotid arteries of the dog (Berne et al. 1974). Potential interventions on ABs would be particularly challenging, mostly due to their highly heterogeneous anatomy and close proximity to the largest vessels of the mediastinum (O'Regan & Majcherczyk, 1982; Kumar & Prabhakar, 2012).

Figure 3. Changes in hypoxic responses following bilateral carotid body resection.

^* P < 0.05 compared to baseline; note virtually complete elimination of ventilatory response (HVR, filled bars), significant decrease in systolic blood pressure response (SBP, open bars) and no change in heart rate response (HR, shaded bars) one month after bilateral carotid body removal. Based on Niewinski et al. 2014 a.

Loss of hypoxic sensing

The most concerning issue pertinent to CB intervention is the reduction or loss of hypoxic sensing. In patients with systolic HF undergoing bilateral CB resection we documented virtually complete elimination of hypoxic ventilatory response (Niewinski et al. 2014 a; see Fig. 3). Out of > 15000 cases of CB resection performed historically only one case of death possibly related to the loss of hypoxic sensing has been reported (Paton et al. 2013). Notably, as the majority of trials of CB resection for the treatment of respiratory disorders was carried out in the 1960 s and 70 s, a detailed follow‐up including oxygenation and sleep studies might not have been always performed. As central sensors detecting $P_{\mathrm{aC}{\mathrm{O}}_2}$ changes are delayed in relation to the CBs, bilateral CB ablation may result in prolonged apnoeas, leading to deeper desaturations at night. This effect might be particularly prominent in individuals with pre‐existing sleep disordered breathing (especially of the obstructive type) and in some cases may lead to serious clinical consequences (Gami et al. 2013). On the other hand, unilateral CB resection might have a less substantial effect on nocturnal oxygenation, possibly because of the preservation of hypoxic sensing by the contralateral CB.

Changes in arterial blood gas analysis might also be expected following CB intervention, including a decrease in $P_{\mathrm{a}{\mathrm{O}}_2}$ and a rise in $P_{\mathrm{aC}{\mathrm{O}}_2}$ – particularly in the group of bilateral CB ablation. These changes could be mediated by the reduction in resting ventilation caused by the loss of excitatory input from the CB to the central chemoreceptors (Nattie & Li, 2012).

Summary

In conclusion, CB modulation may constitute a novel and promising treatment modality in systolic HF. Based on the available data, CB ablation can be expected to decrease sympathetic tone and improve exercise tolerance of HF patients. The prognostic benefits of CB modulation cannot be predicted at this moment in time. Several potential obstacles might, however, complicate CB oriented interventions. Of these, the loss of hypoxic sensing is the most concerning one. It can potentially lead to significant desaturation at night with associated clinical consequences. The results of ongoing trials of unilateral and bilateral surgical carotid body removal (NCT01782677, NCT01653821) will possibly answer at least some of the questions raised in this review.

Additional information

Competing interests

P.N. has received research support from CIBIEM Inc., a company with an interest in carotid body modulation.

Biography

Piotr Niewinski is clinical scientist and active cardiologist specializing in heart failure and invasive electrophysiology. He received his PhD from Wroclaw Medical University in 2015 for his research focused on pathophysiology of peripheral chemoreceptors in systolic heart failure. He was involved in first‐in‐man trials of carotid body modulation. He is a member of the Polish Society of Cardiology and European Society of Cardiology.