Background Noninvasive ventilation (NIV) represents an effective treatment for chronic respiratory failure. However, empirically determined NIV settings may not achieve optimal ventilatory support. Therefore, the efficacy of NIV should be systematically monitored. The minimal recommended monitoring strategy includes clinical assessment, arterial blood gases (ABG) and nocturnal transcutaneous pulsed oxygen saturation (SpO 2 ). Polysomnography is a theoretical gold standard but is not routinely available in many centers. Simple tools such as transcutaneous capnography (TcPCO 2 ) or ventilator built-in software provide reliable informations but their role in NIV monitoring has yet to be defined. The aim of our work was to compare the accuracy of different combinations of tests to assess NIV efficacy. Methods This retrospective comparative study evaluated the efficacy of NIV in consecutive patients through four strategies (A, B, C and D) using four different tools in various combinations. These tools included morning ABG, nocturnal SpO 2 , TcPCO 2 and data provided by built-in software via a dedicated module. Strategy A (ABG + nocturnal SpO 2 ), B (nocturnal SpO 2 + TcPCO 2 ) and C (TcPCO 2 + builtin software) were compared to strategy D, which combined all four tools (NIV was appropriate if all four tools were normal). Results NIV was appropriate in only 29 of the 100 included patients. Strategy A considered 53 patients as appropriately ventilated. Strategy B considered 48 patients as appropriately ventilated. Strategy C misclassified only 6 patients with daytime hypercapnia. Conclusion Monitoring ABG and nocturnal SpO 2 is not enough to assess NIV efficacy. Combining data from ventilator built-in software and TcPCO 2 seems to represent the best strategy to detect poor NIV efficacy. Trial registration Institutional Review Board of the Société de Pneumologie de Langue Française (CEPRO 2016 Georges)