One of the major causes of emergency admissions is acute respiratory failure (ARF) [1]. Noninvasive ventilation (NIV) is increasingly used for ARF management because of its efficacy (particularly in hypercapnic patients), achievable goals and because it can be outside the intensive care unit [2]. The benefits of NIV are still under investigation in various hypoxaemic respiratory failure (HRF) aetiologies, such as asthma, pneumonia, immunosuppression and acute respiratory distress syndrome [1]. We read with great interest the study presented by HUKINS et al. [3], which for the first time, successfully displays a dose-response relationship between the actual mask-on usage period and hospital survival in ARF patients. Notably, the authors took great efforts to compare their results with previous trials, which mainly investigated the total period of NIV therapy, including both mask-on and mask-off times. The study also provides beneficial data regarding the survival rate and its relationship with therapy intensity calculated as a proportion of mask-on usage from a total NIV period, as well as the association of survival with concomitant patient comorbidities. The large sample size (n=654) collected over 5 years was considered a major advantage of this retrospective study, as it increased the applicability of its findings, especially in patients with hypercapnic respiratory failure (HCRF), who represented ∼91% of the included patients. However, we feel that with regard to the methodology, some key aspects require clarification. First, HCRF and HRF groups are not matched so the study outcomes are mainly based on the HCRF group, with a ceiling effect at 24 h of cumulative mask-on usage. The reliability of the results in hypoxaemic patients is therefore limited and further investigations are needed.