Background: Whether changes in the arterial carbon dioxide tension (PaCO 2 ) during long-term non-invasive positive pressure ventilation (NPPV) are related to a good prognosis remains unclear. Methods: Data on 190 patients with restrictive thoracic disease who received long-term NPPV were studied retrospectively. The annual changes in PaCO 2 during NPPV were determined with a simple linear regression method in each patient who had at least four 6-month intervals of PaCO2 data. The annual changes in the PaCO 2 during long-term NPPV and possible confounders were analyzed, with discontinuation of long-term NPPV as the main outcome. Results: One hundred twenty-five patients had more than four 6-month intervals of PaCO 2 data and were performed further studied. PaCO 2 during long-term NPPV decreased in 41 patients (Group 1; less than 0 mmHg/y), increased slightly in 42 patients (Group 2; between 0 and 1.85 mmHg/y), and increased greatly in 42 patients (Group 3; more than 1.85 mmHg/y). Lower annual changes in PaCO 2 (p=0.0002) and a control ventilator mode (p=0.008) were associated with a significantly better probability of carrying on NPPV, while lower PaCO 2 3-6 months after start of long-term NPPV was not (p=0.11). 10-year probabilities of continuing NPPV were 69% for Group 1, 39% for Group 2, and 12% for Group 3. Conclusion: Lower annual change in PaCO 2 during long-term NPPV was significant prognostic variables.
Tsuboi et al. (Sun,) studied this question.