Association of low PaCO2 with central sleep apnea and ventricular arrhythmias in ambulatory patients with stable heart failure.

BACKGROUND Central sleep apnea frequently occurs in patients with heart failure. Because it is not practical to perform sleep studies on all patients, readily available laboratory tests that predict sleep apnea would be clinically useful. Arterial PCO2 has a profound influence on breathing during sleep: When it decreases below a certain threshold, apnea occurs. OBJECTIVE To study the value of a low PaCO2 while patients are awake in predicting central sleep apnea in patients with stable, treated heart failure. DESIGN Prospective study. SETTING Referral sleep laboratory of a Department of Veterans Affairs Medical Center. PARTICIPANTS 59 patients with left ventricular ejection fractions of 45% or less. MEASUREMENTS Arterial blood gases and hydrogen ion concentrations were measured, and cardiac radionuclide ventriculography, Holter monitoring, and polysomnography were done. RESULTS Patients were classified as eucapnic (PaCO2 > 35 and < 44 mm Hg [n = 41]) or hypocapnic (PaCO2 < or = 35 mm Hg [n = 18]). The mean (+/- SD) hourly episodes of apnea or hypopnea (36 +/- 25 and 20 +/- 27; P = 0.015), the prevalence of central sleep apnea (78% and 39%; P = 0.01), and the mean hourly occurrences of ventricular tachycardia (2 +/- 3 and 0.1 +/- 0.1; P = 0.003) were significantly greater in hypocapnic patients than in eucapnic patients. CONCLUSION Data on patients with heart failure in this study are consistent with the physiologic notion that a low PaCO2 results in ventilatory instability and central apnea during sleep. The positive predictive value of a low PaCO2 for central sleep apnea is 78%. The prevalence of ventricular tachycardia was 20 times greater in hypocapnic patients than in eucapnic patients.