A 57-yr-old man with idiopathic central apnea is reported. He presented at our hospital complaining of excessive daytime sleepiness. Polysomnography, including esophageal pressure monitoring, confirmed central sleep apnea with an apnea index of 27/h. He had mild non-insulin-dependent diabetes mellitus (NIDDM) but no signs of diabetic neuropathy or other background diseases. The ventilatory responses to hypoxia and hypercapnia tested while he was awake indicated increased respiratory chemosensitivity. We applied nasal continuous positive airway pressure (CPAP) and bilevel positive airway pressure (BPAP) in an attempt to compare the possible difference in therapeutic efficacy. Although nasal CPAP completely reversed central apnea, nasal BPAP adversely affected both apnea length and frequency in an applied pressure-dependent manner. Arterial blood gas analyses while he was being treated indicted alveolar hypoventilation with CPAP and hyperventilation with BPAP. Additionally, administration of a mixed gas containing 5% CO2 through a face mask had a significant effect on the disappearance of central apnea in this patient. These findings support the theory that the arterial PCO2 level is critical in generating idiopathic central apnea and that nasal CPAP therapy may be effective in eliminating central apnea by raising the PaCO2.