Treating chemoreflex in heart failure: modulation or demolition?

Dysregulation of baro-, ergo- and chemoreflex in patients with heart failure (HF) elicits autonomic imbalance associated with parasympathetic withdrawal and increased adrenergic drive to the heart, and contributes to worsening arrhythmic burden and haemodynamics. Hence, on top of guideline-recommended medical therapy, all visceral feedbacks have been considered in HF patients as additional targets; baroreflex activation therapy for baroreceptors (Sabbah et al. 2011), physical training for muscle metaboreceptors (Piepoli et al. 1996), and, recently, carotid body (CB) denervation for chemoreceptors (Marcus et al. 2014). In particular, increased peripheral (PC) and central chemoreceptor (CC) sensitivity is frequently observed in HF patients. This phenomenon, is recognized as an independent prognostic marker in HF (Ponikowski et al. 2001; Giannoni et al. 2009), being associated with autonomic imbalance, adrenergic activation, and with ventilatory instability (Cheyne–Stokes respiration); therefore, it has been proposed as a therapeutic target (Paton et al. 2013). Schultz and colleagues have investigated the causal link between HF and PC sensitivity. They first showed an increase in PC sensitivity (Sun et al. 1999) in a rabbit pacing-induced HF model, probably due to reduced CB perfusion and suppression of outward voltage-gated K+ currents, deriving from altered carbon monoxide/nitric oxide and angiotensin-II signalling in glomus cells (Schultz & Marcus, 2012). They also showed that an increase in PC activity leads to adrenergic activation in normoxia, reversible after pure oxygen inhalation, and to a higher sympathetic response to hypoxia (Sun et al. 1999). Finally, they showed the reversibility, after bilateral cryoablation of CB, of the several negative correlates of overactive PCs, again in the rabbit pacing-induced HF model (Marcus et al. 2014). Indeed, bilateral CB ablation was found to decrease baseline ventilation and the hypoxic ventilatory response to a level lower than in sham animals. The baseline renal sympathetic nerve activity (RSNA) and the hypoxia-related RSNA response showed a similar behaviour after CB ablation. These effects on adrenergic outflow and ventilation were followed by a decrease in arrhythmias and ventilatory instability. Finally, although no effect on left ventricular ejection fraction was reported, reduced diastolic and systolic volumes were found in CB-ablated compared to sham HF rabbits. The paper by Marcus et al. (2014) is indeed an interesting advance in research on feedback control, though it raises some methodological and conceptual questions. First, ventricular pacing used to produce HF could potentially increase the arrhythmic risk (Gardiwal et al. 2008), leading to the overestimation of the positive effect of CB ablation on rhythm stability. More importantly, pacing takes up to 4 weeks to induce HF, so it should be considered as a subacute model of HF rather than a chronic model, able to induce haemodynamic compromise, sensed by PCs (Schultz & Marcus, 2012), but less likely to be associated with significant neurohormonal activation and with the global feedback resetting seen in the chronic scenario. This caveat may hamper the direct extension of the findings by Marcus and colleagues to the clinical setting of chronic HF. Second, PCs are commonly considered the only sensors of hypoxia available in the body, and also an ‘emergency valve’, quickly activated in conditions of asphyxia, with hypoxia and hypercapnia simultaneously present. Although both carotid and aortic bodies function as PCs, with interspecies differences, CBs act as the main hypoxia sensors in humans (Prabhakar & Peng, 2004). Moreover, in contrast to animal studies, failure to restore the hypoxic ventilatory response from aortic bodies after long-term CB removal has been described in humans with asthma (Prabhakar & Peng, 2004). Therefore, an irreversible bilateral surgical CB demolition seems far from desirable, especially in HF patients, when in acute life-threatening conditions, such as, pulmonary oedema or cardiogenic shock, the adrenergic and ventilatory response linked to hypoxia may instead be necessary. Bilateral CB ablation may also increase arterial carbon dioxide by reducing baseline ventilation, and reduce pH (unfortunately not measured in the experiments of Marcus et al. 2014), which seems undesirable, too, since maintenance of pH homeostasis and oxygen availability at the tissue level is the primary purpose of the chemoreflex system (Blain et al. 2010). In conclusion, monolateral CB surgery, as already tested in mice (Del Rio et al. 2013), or, alternatively, a reversible pharmacological/non-pharmacological CB modulation could succeed in blunting the negative effects of PC overactivity, without affecting its physiological function, with a net long-term positive effect on survival.