Mouth leak is common during nasal ventilatory assistance, but its effects on ventilatory support and on sleep architecture are unknown. The acute effect of sealing the mouth on sleep architecture and transcutaneous carbon dioxide tension (Ptc,CO 2 ) was tested in 9 patients (7 hypercapnic) on longterm nasal bilevel ventilation with symptomatic mouth leak.Patients slept with nasal bilevel ventilation at their usual settings on two nights in random order. On one night, the mouth was taped closed. Nocturnal nasal bilevel ventilatory support has become an important and successful therapy for respiratory failure during sleep [1±3]. At present, nasal masks are more common than mouthpieces, mouth-nose masks or full face masks (perhaps because of issues related to comfort, fit and deadspace) but a major practical problem is escape of air via the mouth, as has been well documented for nasal continuous positive airway pressure (CPAP) [4,5].In the only study to date to directly measure the effect of mouth leaks on effectiveness of ventilatory support, CAR-REY et al.[6] investigated the effect of voluntary mouth opening on diaphragm electromyogram (EMG) activity during awake nasal intermittent positive pressure ventilation in 5 subjects. With the mouth closed, diaphragm EMG dropped to 15% of unassisted control, but with the mouth open, returned to 98% of unassisted control, implying complete loss of ventilatory support.Mouth leak is ubiquitous during noninvasive ventilatory support in sleep. BACH et al. [7] demonstrated that during nasal intermittent positive pressure ventilation (NIPPV), severe leak (>33% of tidal volume escaping) was present for a median of 55% of sleep time, and was associated with 7.5 desaturations (of at least 4%) per hour of sleep.MEYER et al. [8] found severe mouth leak during most of sleep time and 100% of slow wave sleep.Mouth leak could lead to severe sleep fragmentation, either directly via airway irritation, or indirectly via a reduction in effectiveness of ventilation as discussed above. In the study by BACH et al. [7], there were a total of 34 arousals . h -1 and 74% of desaturations were terminated by either an arousal or a lightening of sleep stage. The authors concluded that leak was reducing ventilatory effectiveness and causing desaturation, and that arousal reversed the leak and desaturation. Similarly, in the study of MEYER et al. [8] there were 46 arousals . h -1 in stages I-II nonrapid eye movement (NREM) sleep, associated in time with mouth leak. However, association in time does not establish causality, and it is not known whether preventing the mouth leak would improve sleep quality.The purpose of the present paper was to directly assess the effect of mouth leak on sleep architecture and transcutaneous carbon dioxide tension (Ptc,CO 2 ) during nasal bilevel ventilatory support, by preventing the leak. To do this, the authors measured the first-night effect of taping the mouth closed, on sleep architecture and Ptc,CO 2 in patients already on long-term nasal bilevel venti...