The importance of controlling PaCO₂ throughout long-term noninvasive ventilation.

Background: The significance of changes in P(aCO2) during long-term noninvasive ventilation (NIV) on prognosis remains unclear. We aimed to clarify whether stabilizing P(aCO2) during NIV had a favorable prognostic effect.

Methods: Data from 190 subjects with restrictive thoracic disease and who received long-term NIV were studied retrospectively. The annual change in P(aCO2) during NIV was determined using a simple linear regression method for each subject who had at least 4 6-month intervals of P(aCO2) data. Annual changes in P(aCO2) during long-term NIV and possible confounders were analyzed with discontinuation of long-term NIV as the main outcome.

Results: One hundred and twenty-five subjects who had > 4 6-month intervals of P(aCO2) data were included in the study. P(aCO2) during long-term NIV decreased in 41 subjects (group 1; < 0 mm Hg/y), increased slightly in 42 subjects (group 2; between 0 and 1.85 mm Hg/y), and increased significantly in 42 subjects (group 3; > 1.85 mm Hg/y). Smaller annual changes in P(aCO2) (P < .001) and a control ventilator mode (P = .008) were associated with a significantly higher probability of continuing NIV, compared with decreased P(aCO2) 3-6 months after the start of long-term NIV (P = .11). The 10-y probability of continuing NIV was 69% in group 1, 39% in group 2, and 12% in group 3.

Conclusions: A decrease in the annual change of P(aCO2) during long-term NIV was shown to be a significantly prognostically favorable factor. Efforts to reduce P(aCO2) should be made if P(aCO2) increases at a greater rate during long-term NIV.