Adapting Our Approach to Treatment-Emergent Central Sleep Apnea

While obstructive sleep apnea is effectively treated with continuous positive airway pressure (CPAP), roughly 10% of initial CPAP titrations result in emergence of central apneas or hypopneas,^1–3 a phenomenon referred to as treatment-emergent central sleep apnea (CSA). Although the majority of treatment-emergent CSA resolves with ongoing CPAP therapy,⁴ some patients exhibit persistent CSA, creating a therapeutic challenge.

Currently, the treatment options are to either continue using CPAP assessing whether central apneas naturally resolve (as occurs in > 90% of patients by 3 months^2,4–7) or try changing the mode of positive airway pressure (PAP) to either bilevel PAP (BPAP) with a backup rate or adaptive servoventilation (ASV). Although data are sparse, limited available evidence suggests that both BPAP and ASV improve CSA, although ASV may have a slight advantage over BPAP.^8,9 However, there has been a lack of longitudinal data examining the efficacy of either of these two modes of PAP. In this edition of SLEEP, Dellweg and colleagues¹⁰ take the first step in filling the research void by conducting a randomized prospective comparative effectiveness study in patients whose CSA persisted after 6 weeks of CPAP. The study showed that both ASV and BPAP effectively resolve treatment-emergent CSA during the titration night, but only ASV maintained an effective suppression of CSA over time. Thus, ASV appears superior to BPAP for the long-term treatment of persistent CSA.

The most intriguing observation from Dellweg et al.¹⁰ was the re-emergence of central events after 6 weeks with BPAP, despite the apparent absence of events on the titration night. Ideally, BPAP should effectively eliminate all central events if there is an adequate backup rate, tightly sealed mask, sufficient pressure support (inspiratory driving pressure), and stable respiratory mechanics. Yet CSA elimination is not always seen in practice for reasons that are unclear.

What might explain the occurrence of CSA on BPAP therapy? When BPAP is delivered noninvasively via a mask, the ultimate effect on tidal volume is determined by the respiratory system mechanics and the level of leak present. Importantly, by augmenting tidal volume, BPAP reduces CO₂ and thus the intrinsic ventilatory drive.¹¹ If CO₂ is sufficiently lowered, apnea will be produced if BPAP fails to deliver tidal volume effectively. One way by which BPAP may fail to deliver this volume is if the airway is closed. Airway closure can be observed during central apnea either via active laryngeal closure reflexes induced by hypocapnia¹² or by loss of chemical drive to upper airway muscles in those mechanically predisposed to pharyngeal airway occlusion.^13–15 A variable mask leak could also be expected to contribute to CSA either via washing out anatomical dead-space and thereby contributing to the hypocapnia or by simply impairing the ability of the device to deliver effective breaths. Furthermore, in a background of hypocapnia, any failure to detect apnea and trigger the machine to deliver a breath with appropriate timing will also allow CSA to occur. In addition, BPAP may promote CSA in the presence of a variable upper airway resistance (e.g., during inadequate expiratory PAP), which should, in principle, act to destabilize the ventilatory control system. Likewise, state instability (e.g., as occurs with a low respiratory arousal threshold) may also promote CSA since a change in sleep state can be accompanied by large changes in ventilatory drive.¹⁶ State transition apneas may further be promoted with arousal-induced reductions in PaCO₂ levels.¹⁷ This mechanism may be particularly important if patients are uncomfortable or arousable while using the machine.

An unresolved question in the report of Dellweg et al.¹⁰ concerns why central events would be absent on the BPAP titration night and redevelop over time. Several possibilities come to mind. (1) As sleep was not assessed on the titration night but was assessed at follow-up, it is possible that more stable breathing was seen on titration as a consequence of long periods of undocumented wakefulness. (2) Patient-BPAP synchrony may have improved over time, leading to greater augmentation of tidal volumes, larger reductions in CO₂, and thereby increasing the likelihood of central events. (3) If the effectiveness of pressure support was reduced over time (e.g., increased mask leak), the effective settings used on the titration night may have become ineffective later. (4) The reduction in the arousal threshold with sleep apnea treatment^18–20 could make arousal and state instability more prominent. (5) Finally, as Dellweg et al.¹⁰ suggest, a deleterious long-term adaptation to BPAP-induced hyperventilation/hypocarbia may promote susceptibility to CSA. Regardless of the mechanism responsible, the re-emergence of CSA would not be expected to occur with ASV, since any oscillations in ventilatory output are cleverly monitored to allow pressure support to be minimized when it is not needed (i.e., during hyperventilation) and maximized when it is most needed (i.e., during hypoventilation). If the finding of re-emergent central events on BPAP is confirmed in future studies, elucidating the mechanisms may provide for further improvements to BPAP utilization/technological adjustments (e.g., titration/triggering) and/or enable selection of appropriate patients likely to benefit from this form of therapy.

Based on the new and interesting results,¹⁰ a number of key questions emerge for scientists. (1) Could inadequate expiratory PAP contribute to dynamic instability of the upper airway promoting unstable ventilatory control?²¹ (2) In causing treatment-emergent CSA, does CPAP therapy simply unmask an inherent ventilatory instability or does it contribute to destablizing the ventilatory control system? (3) Does the lowering of the arousal threshold with apnea therapy^15,17,18 contribute to state instability and CSA reappearance?⁴ The findings of Dellweg et al.¹⁰ also raise important questions for clinicians. (1) Does a poor initial experience with PAP in patients who develop emergent central apneas contribute to poor adherence and if so, does early intervention improve outcome in these patients?²² (2) Could adjustments/improvements in BPAP settings and technology be used to provide similar results to those with ASV in selected patients? (3) Could adjustment of the arousal threshold (e.g. using sedative/hypnotic agents²³) improve the state transition apneas which may develop over time? These questions clearly highlight the need for further efforts dedicated towards elucidating underlying mechanisms as well as determining the impact of interventions on patient-centered outcomes in patients with treatment-emergent CSA.

CITATION

Edwards BA; Malhotra A; Sands SA. Adapting our approach to treatment-emergent central sleep apnea. SLEEP 2013;36(8):1121–1122.

DISCLOSURE STATEMENT

Drs. Edwards and Sands have no conflicts of interest to disclose. Dr. Malhotra was a consultant for Philips Respironics, SHC, SGS, Apnex Medical, Pfizer, Apnicure, but has relinquished all outside personal income since May 2012.