Mechanism of central sleep apnea resolution via carbon dioxide stimulation in patients with heart failure.

Abstract

Rationale: Patients with heart failure commonly exhibit central sleep apnea (Cheyne-Stokes respiration, CSR), which can adversely affect sleep, sympathetic activity and cardiac function. Unfortunately, treatment with continuous positive airway pressure is not universally effective or tolerated. An alternative approach, employing respiratory stimulants such as supplemental inspired CO2, consistently resolves the ventilatory oscillations of CSR for reasons not fully elucidated. Objective(s): Here we critically-examine the idea that control theory quantitatively explains CSR resolution with inspired CO2. Specifically, CSR is the consequence of "loop gain" being in excess of the critical threshold for instability (1.0). We test two hypotheses: 1) Inspired CO2 stimulation powerfully reduces loop gain by lowering "plant gain" (the effectiveness of swings in ventilation on alveolar PCO2) via narrowing the alveolar-inspired CO2 difference; 2) Inspired CO2 stabilizes breathing when loop gain can be reduced below 1.0. Method(s): In 12 patients with heart failure, we administered 1%, 2% or 3% CO2 during N = 95 episodes of CSR. To determine the stabilizing mechanism of CO2, we measured plant gain (DELTAPETCO2/DELTAventilation) and controller gain (chemoreflex: DELTAventilation/DELTAPETCO2) by fitting models to measured data. Loop gain at baseline (LGbaseline) was calculated from the ventilatory pattern of CSR (ventilatory duration / cycle duration) and used to predict CSR resolution (LGpost < 1) based on the expected reduction in plant gain ('m'), where LGpost = LGbaseline[100-m x FICO2]/100. Result(s): Inspired CO2 lowered the alveolar-inspired CO2 gradient by m = 18% for each 1% rise in inspired CO2 (during wake and during CSR), and lowered plant gain in direct proportion (y = [0.94 +/- 0.05]x, p < 0.0001) as expected by theory; controller gain was unchanged. Loop gain at baseline was higher prior to failed versus successful 1% and 2% interventions (p < 0.01). When loop gain was predicted to remain above 1 after CO2 treatment (LGpost > 1) CSR persisted on 23/25 occasions, whereas lowering loop gain to below 0.8 resolved CSR on 37/37 occasions. Conclusion(s): Inspired CO2 abolishes the ventilatory oscillations of CSR in patients with heart failure by a powerful reduction of plant gain. Our analysis shows that loop gain quantitatively determines the level of respiratory stimulation sufficient to stabilize breathing.