BACKGROUND AND OBJECTIVES Non-invasive positive pressure ventilation (NPPV) is an established treatment in restrictive thoracic disorders (RTD) with chronic hypercapnic respiratory failure. The aim of this study was to identify predictors of long-term survival for patients on NPPV therapy. METHODS In a 10-year retrospective cohort of patients with RTD and chronic hypercapnic respiratory failure, survival and the predictive value of nocturnal and daytime blood gases, lung function and laboratory data measured before initiation of NPPV were assessed. The impact of ventilator settings and daily use of NPPV on survival were also evaluated. Patients were re-admitted every 6 months for follow-up assessment. RESULTS The study recruited 77 patients; 18 died during the study period and three ceased NPPV. Respiratory failure caused eight of the nine respiratory deaths (88.9%). One-, 2- and 5-year survival rates were 92.5%, 81.0% and 59.0%, respectively. In univariate analyses, higher night-time PaCO(2), base excess (night- and daytime) and lower Hb at baseline were associated with significantly worse survival (P < 0.05). Multivariate Cox regression analysis revealed night-time PaCO(2) as an independent predictor of survival (P = 0.042). The small differences in daily duration of use of NPPV and ventilator settings were not significantly related to survival. At follow up, significant improvements were observed for blood gases, lung and respiratory muscle function, as well as a decrease in Hb level (P < 0.01 each). CONCLUSIONS Base excess, Hb and particularly nocturnal PaCO(2) are relevant prognostic factors for survival in RTD and should be considered in assessing patients receiving NPPV.