Exercise End-Tidal CO2 Predicts Central Sleep Apnea in Patients With Heart Failure

Abstract

Increased CO2 chemosensitivity and augmented exercise ventilation are characteristic of patients with heart failure (HF) with central sleep apnea (CSA). The aim of this study was to test the hypothesis that decreased end-tidal CO2 by cardiopulmonary exercise testing predicts CSA in patients with HF. Consecutive ambulatory patients with New York Heart Association II to III HF were prospectively evaluated by CO2 chemosensitivity by rebreathe, cardiopulmonary exercise testing, and polysomnography (PSG). Subjects were classified as having either CSA (n = 20) or no sleep apnea (n = 13) by PSG; a central apnea-hypopnea index (AHI) ≥ 5 was used to define CSA. Subgroups were compared by t test or Mann-Whitney test and data summarized as mean ± SD. P < .05 was considered significant. At rest, subjects with CSA had higher central CO2 chemosensitivity (Δminute ventilation [V.e]/Δpartial pressure of end-tidal CO2 [Petco2], 2.3 ± 1.0 L/min/mm Hg vs 1.6 ± 0.4 L/min/mm Hg, P = .02) and V.e (15 ± 7 L/min vs 10 ± 3 L/min, P = .02) and lower Petco2 (31 ± 4 mm Hg vs 35 ± 4 mm Hg, P < .01) than control subjects. At peak exercise, the ventilatory equivalents per expired CO2 (V.e/V.co2) was higher (43 ± 9 vs 33 ± 6, P < .01) and Petco2 lower (29 ± 6 mm Hg vs 36 ± 5 mm Hg, P < .01) in subjects with CSA. In addition, CO2 chemosensitivity, peak exercise V.e/V.co2, and Petco2 were independently correlated with CSA severity as quantified by the AHI (P < .05). Peak exercise Petco2 was most strongly associated with CSA (OR, 1.29; 95% CI, 1.08-1.54; P = .01; area under the curve, 0.88). In patients with HF and CSA, ventilatory drive is increased while awake at rest and during exercise and associated with heightened CO2 chemosensitivity and decreased arterial CO2 set point.