The precise mechanism of dyspnea in chronic heart failure (CHF) is not known. We studied the chemosensitivity of 14 CHF patients (11 men; age 62 ± 1 [mean ± SEM] years; radionuclide left ventricular ejection fraction 29 ± 4) and 9 controls (6 men; age 56 ± 3 years) by assessing the ventilatory response to hypoxia using transient pure nitrogen inhalation and to hypercapnia using single breath inhalation of 13% carbon dioxide in air The ventilatory response to hypoxia was significantly increased in CHF compared with normal subjects (0.67 ± 0.14 vs 0.24 ± 0.05 I/min/% SaO 2 , P = 0.013) but not to hypercapnia (0.31 ± 0.03 vs 0.26 ± 0.05 I/min/mmHg CO 2 , P = NS). Mean maximal oxygen consumption on cardiopulmonary exercise testing was 16.4 ± 0.8 vs 26.7 ± 1.8 ml/min/kg ( P < 0.01) and the slope relating minute ventilation (V E ) and carbon dioxide production (VCO 2 ) was 34.1 ± 2.0 vs 25.8 ± 0.8 ( P < 0.01). There was a correlation between the hypoxic ventilatory response and V E /VCO 2 slope ( r = 0.43, P = 0.043) and an inverse trend between this response and maximal oxygen consumption (r = -0.36, P = 0.099). There was no correlation between the hypoxic ventilatory response and ejection fraction ( r = -0.25, P = NS). We conclude that there is an increased chemosensitivity to hypoxia in CHF and this may contribute to the mechanism of dyspnea in this condition.