Patients with neuromuscular disease (NMD) who require long-term ventilation normally have the ventilation set using empirical daytime parameters. We evaluated arterial blood gases (ABG), breathing pattern, respiratory muscle function, and sleep architecture during ventilation with two noninvasive Pressure Support Ventilation (nPSV) settings in nine patients with NMD. The two settings were randomly applied: the usual (US), with the nPSV setting titrated on simple clinical parameters, and the physiological (PHYS), tailored to the patient's respiratory effort. During wakefulness, nPSV significantly improved ABG and minute ventilation and reduced the diaphragmatic pressure-time product (PTPdi/breath), independently of the type of setting (PTPdi/breath spontaneous breathing 5.7 +/- 2.4, US 3.2 +/- 2, PHYS 3.6 +/- 1.6 cm H2O . seconds(-1), p < 0.001). However, during sleep, PHY nPSV resulted in a significant improvement of gas exchange, sleep efficiency (71.7% +/- 14 US vs. 80.6% +/- 8.3 PHYS, p < 0.01) and % of REM sleep (9.1% +/- 7 US vs. 17.3% +/- 5.4 PHYS, p < 0.01). This improvement was significantly correlated with the reduction in ineffective efforts. In NMD, nPSV is effective in improving daytime ABG and in unloading inspiratory muscles independently of whether it is set on the basis of the patient's comfort or the patient's respiratory mechanics. However, PHYS was associated with better sleep architecture and nighttime gas exchange.

Using a simple decisional flowchart we were able to remove tracheotomy cannula in almost 80% of the patients with spontaneous breathing autonomy without major clinical complications. Further larger prospective studies are needed to confirm this clinical approach in larger and different populations.

We conclude that, over time, experience with NIV may progressively allow more severely ill patients to be treated without changing the rate of success. The daily cost of NIV per patient can be reduced by treating less severely ill patients outside the RICU.