Sleep and respiratory function after withdrawal of noninvasive ventilation in patients with chronic respiratory failure.

BACKGROUND In patients with restrictive thoracic disease, little is known about changes in sleep and breathing if the patient stops using nocturnal noninvasive ventilation (NIV). Better understanding of those changes may affect NIV management and improve our understanding of the relationship of night-to-night variability of respiratory and sleep variables and morning gas exchange. METHODS With 6 stable patients with restrictive chronic respiratory failure who were being treated with home NIV we conducted a 5-step study: (1) The subject underwent an in-hospital baseline sleep study while on NIV, then next-morning pulmonary function tests. (2) At home, on consecutive nights, the subject underwent the same sleep-study measurements while not using NIV, until the patient had what we defined as respiratory decompensation (oxygen saturation measured via pulse oximetry [S(pO(2))] < 88% or end-tidal CO(2) pressure [P(ETCO(2))] > 50 mm Hg, with or without headaches, fatigue, or worsening dyspnea). Each morning after each home sleep-study night off NIV, we also measured S(pO(2)) and P(ETCO(2)). (3) The patient returned to the hospital for a second overnight assessment, the same as the baseline assessment except without NIV. (4) The patient went home and restarted using NIV with his or her pre-study NIV settings. (5) After the number of nights back on home NIV matched the number of nights the patient had been off NIV, the patient returned to the hospital for a third in-hospital assessment. We measured static lung volumes, maximum inspiratory and expiratory static mouth pressure, breathing pattern, arterial blood gases, S(pO(2)), P(ETCO(2)), and full overnight polysomnography values. RESULTS Respiratory decompensation occurred 4-15 days after NIV discontinuation (mean 6.8 d). On the first and second in-hospital assessment nights, respectively, the mean nadir nocturnal S(pO(2)) values were 84 +/- 2% and 64 +/- 4%, the total apnea-hypopnea index values were 0 +/- 0 and 9 +/- 2, and the obstructive hypopnea index values were 0 +/- 0 and 7 +/- 1 episodes per total sleep hour. Respiratory events started on the first night off NIV. Spirometry, muscle strength, and sleep architecture did not change significantly. With resumption of NIV, baseline conditions were recovered. CONCLUSIONS NIV discontinuation in patients with restrictive chronic respiratory failure previously stabilized on NIV promptly leads to nocturnal respiratory failure and within days to diurnal respiratory failure. Stopping NIV for more than a day or two is not recommended.