Central sleep apnea: treating the epiphenomenon and neglecting the disease?

Central sleep apnea (CSA) has always been a challenge to diagnose and treat. Especially patients with fluid overload, such as patients with heart failure (HF) present with CSA [1]. In patients with HF, CSA commonly occurs in the form of hunter-cheyne-stokes respiration (CSR-CSA), a pattern of periodic breathing with recurring cycles of prolonged crescendo-decrescendo hyperpnea that culminates in an apnea or hypopnea. Pulmonary congestion, increased central and peripheral chemosensitivity, and oscillation of the PaCO₂ around the apneic threshold appear to be key factors in CSA development and perpetuation, specifically in patients with HF [2]. CSA is associated with repeated episodes of intermittent hypoxia, and arousals from sleep throughout the night, leading to elevated sympathetic activity [3, 4].

A key deleterious effect of CSA in HF patients is the occurrence of repeated bursts of sympathetic activity with each central event [5]. The alternance of apneas and hyperventilation phases leads to sympathovagal imbalance [6]. Specifically, CSR-CSA has been linked with greater urinary and plasma norepinephrine levels, increased norepinephrine spillover, and muscle sympathetic nerve activity [7]. In patients with CSA bursts of sympathetic activity contribute to oscillations in heart rate and arterial blood pressure and thus cardiac workload [8, 9]. Chronic intermittent hypoxia shortens the atrial effective refractory period and increase the sensitivity to parasympathetic activation and sympathetic potentiation enhancing the vulnerability to atrial arrhythmias [6]. Ultimately, increased sympathetic activity is associated with poor prognosis and higher mortality rates in populations with HF [10].

Heart rate variability (HRV) is an indirect recognized expression of the autonomous nervous system, with a reduced time domain and spectral power at all frequencies, with increased low-frequency/high-frequency (LF/HF) ratio reflecting impaired parasympathetic control and increased sympathetic dominance [11, 12]. Such abnormalities in HRV have been consistently associated with poor prognosis. In healthy subjects, cardiac vagal activity modulates heart rate responses at high frequencies, whereas the sympathetic nervous system is more slowly adapting and is thus incapable of causing cardiovascular effects at high frequencies [13]. Patients living with HF typically have diminished parasympathetic activity, thus displaying little or no high-frequency spectral power [14]. Although total spectral power of heart rate is attenuated in HF compared with normal subjects, there is a relative increase in spectral power of HRV in the very-low-frequency (VLF) band, corresponding to 0.01–0.04 Hz, which has been associated with CSR-CSA [9, 15].

In the context of HF, interventions that attenuate increased sympathetic activity would therefore reduce cardiovascular mortality linked to HF and CSR-CSA. Taking for granted that therapies targeting the underlying cause, namely HF, would diminish the severity of CSR-CSA and sympathetic activity, interventions directed specifically towards CSA provide heterogeneous results (Table 1) [16, 17].

Table 1.

Effects of Different Treatment Options for Central Sleep Apnea in Patients With Heart Failure on Heart Rate Variability

Article	Patients	Treatment option	Results
Sakakibara 2005 J Cardiol	HF with either CSA and OSA	Nocturnal oxygen	SDNN, LF, and HF↔
Gorbachevski 2020 Int J Cardiol	HFrEF and CSA No heart failure and CSA	ASV and automatic CPAP	HFrEF: ASV and automatic CPAP HRV↔. No heart failure and CSA: CPAP: LF↓ and HF↑, ASV: LF↔ and HF↔
Terziyski 2016 Clin Exp Pharmacol Physiol	HF and CSA	CPAP	SDNN, RMSSD total power, LF, and HF ↓.
D’Elia 2013 J Cardiovasc Med	HFrEF and CSA	ASV	Mean HR ↓. ASV improved SDNN and SDANN compared to baseline
Spießhoefer 2016 J Heart Vessels	HFrHF and CSR-CSA versus control	ASV	ASV increases parasympathetic nervous activity in HF patients
Baumert 2023 Sleep	HFrEF	TPNS	TPNS: VLF↓ and LF↓

Definition of abbreviation: HFrEF, heart failure with reduced ejection fraction; CSR-CSA, Hunter-Cheyne-Stokes respiration with central sleep apnea; CPAP, continuous positive airway pressure; SDNN, Standard deviation of normal–normal interbeat interval; SDANN, normal sinus to normal sinus; RMSSD, root mean square of successive differences; LF, low frequency; HF, high frequency.

The recommended treatment options for CSR-CSA include continuous positive airway pressure, adaptive servo-ventilation (ASV) for patients with a left ventricular ejection fraction ≥ 45% after continuous positive airway pressure failure, and other options such as transvenous phrenic nerve stimulation (TPNS), oxygen and acetazolamide therapy if adequate trials of indicated therapies fail or are not appropriate [18].

The recent published article by Baumert and coauthors [19], addresses the treatment effects of TPNS. TPNS is an implantable device that causes diaphragmatic contraction via unilateral TPNS and has been proposed as an option in selected patients with symptomatic CSA who fail or do not tolerate first choices therapies [20, 21]. The device is programmed to deliver stimulation during sleep and senses respiration via a lead in a thoracic vein [22]. TPNS intervenes by stimulating a respiratory response to a central event and therefore acts by interrupting the trigger of periodic breathing. Arousals from sleep, which can trigger hyperpnea, are ameliorated as well. However, the direct impact of abolishing apnea by TPNS on the failing heart of patients livig with HF have not been tested. Moreover, confirmation that stimulated suppression of CSA reduces sympathetic nerve activation and hyperpnea-induced ventricular arrhythmias in parallel have not been performed yet.

Baumert and coauthors [19], evaluate for the first time the impact of TPNS on the sympathetic nervous system using HRV analysis. TPNS titrated to reduce respiratory events was associated with HRV amelioration in the VLF domain and reduced the normalized low-frequency (LFnu) power during non-REM sleep. However, these results need some interpretation. First, LFnu are more general indicators of aggregate modulation of both the sympathetic and parasympathetic branches of the autonomous nervous systems, rather than a direct expression of sympathetic overactivity [23]. LF/HF is the most widely used HRV index of sympatho-vagal balance between the two branches of the autonomous nervous systems; however, this ratio was not presented [19]. Second, it has been previously demonstrated that in patients with HF heart rate and blood pressure oscillations at a VLF are dependent primarily on periodic oscillations in ventilation, rather than episodic hypoxia [9]. Thus, the resolution of CSA and periodic breathing “per se” might induce VLF amelioration, without meaning a direct improvement of TPNS on the sympathetic overactivity. This being said, only 7 patients out of 22 (32%) presented with HF in the studied cohort and only 17% had CSR-CSA, reducing the impact of VLF results in the studied population.

Although this study is an important step forward in the research on the impact of new CSA treatments, more results regarding the direct impact of TPNS on the cardiovascular system such as blood pressure, cardiac arrhythmias, and cardiac remodeling as well as cardiovascular prognosis need to be performed in patients with HF and preserved ejection fraction. As with other treatment options for patients who fail conventional treatments, TPNS has not yet been compared with other therapies for CSA, and further studies are needed to define optimal patient selection.

Considering the previously described role of sympathetic outflow in the risk of cardiovascular mortality of HF patients with CSA, also the consequences of ASV on the sympathetic overactivation presented contrasting results (Table 1). ASV has first been reported to significantly decrease sympathetic tone and catecholamine levels [24–26]. In other studies, ASV did not modify sympathetic tone evaluated with muscle sympathetic nerve activity [27] and did not improve arrhythmias [28]. However, the latter results are based on the population of a single major study (SERVE-HF). We are waiting for the publication of the results of the ADVENT-HF [29] study to better understand the effects of ASV in patients with HF and reduced ejection fraction and CSA on cardiovascular mortality, arrhythmias, and sympathetic modulation. We cannot expect major randomized long-term trials testing the effects of TPNS in patients with CSA on cardiovascular mortality in the near future. Thus, we may consider effects of surrogates for cardiovascular outcomes such as HRV and/or have to extrapolate from available randomized long-term trials of other treatment options for CSA to guide clinical decision-making.

Treatment of HF has been shown to also abolish CSA [30]. Vice versa the effects of treating CSR-CSA on the main disease (i.e. HF) and on surrogates of cardiovascular damage such as HRV are still heterogeneous and trigger controversial interpretations. Should we go deeper into the pathophysiology [31] of HF with CSA as maybe we are missing something?

Disclosure Statements

Financial disclosure: E.P. has no financial conflicts of interest. M.A. has received consulting fees from ResMed and Philips Respironics, as well as grant support from the ResMed Foundation, Philips Respironics, and the Else-Kroener Fresenius Foundation (2018_A159). Nonfinancial disclosure: authors have no conflicts of interest to disclose.