Introduction Cheyne-stokes Respiration with Central Sleep Apnea (csr-csa) Is a Form of Periodic Breathing in Which Central Apneas and Hypopneas Alter- Nate with Ventilatory Periods Having a Crescendo-decrescendo Pattern of Tidal Vol-

CHEYNE-STOKES RESPIRATION WITH CENTRAL SLEEP APNEA (CSR-CSA) IS A FORM OF PERIODIC BREATHING IN WHICH CENTRAL APNEAS AND HYPOPNEAS ALTERNATE WITH VENTILATORY PERIODS HAVING A CRESCENDO-DECRESCENDO PATTERN OF TIDAL VOLUME. CSR-CSA is common among patients with heart failure (HF), where it is present in 30% to 40% of patients in the largest reported series.1,2 Growing evidence indicates that CSR-CSA is part of a vicious pathophysiologic cycle involving the cardiovascular, pulmonary, and autonomic nervous systems that ultimately contributes to increased mortality among patients with HF.3-5 Ventricular premature beats (VPB) are a risk factor for both arrhythmic and nonarrhythmic death in patients with ischemic heart disease.6 Therefore, one possible explanation for higher mortality in patients with CSR-CSA is increased ventricular irritability due to apnea-related hypoxia, arousals from sleep, sympathetic nervous system activation, and associated elevations in blood pressure and heart rate.3-5,7, 8 A number of previous reports have established an association between CSR-CSA and ventricular ectopy,9,10 but a cause-effect relationship has yet to be definitively established. Notably, Javaheri11 observed, in a posthoc analysis, that those patients whose CSR-CSA was alleviated by application of continuous positive airway pressure (CPAP) also experienced a reduction in the frequency of ventricular ectopy. However, in none of these previous studies was it demonstrated that CSR-CSA actually provokes ventricular ectopy. If CSR-CSA indeed precipitates ventricular ectopy, we hypothesized that VPB would occur more frequently during episodes of CSR-CSA than during periods of regular breathing in the same patients. We further hypothesized that the frequency of VPB during CSR-CSA would be greater during the hyperpneic phase when chemostimulation, blood pressure, and heart rate reach their peak than during the apneic phase when these variables are at their nadir. To test these hypotheses, in patients with HF, we compared the frequency of VPB between CSR-CSA and regular breathing, and between the ventilatory and apneic phases of CSR-CSA.