Introduction Hypoxic challenge test consisting of breathing 15% FiO2 for 20 min with blood gas measurements is recommended by BTS guidelines for the assessment of the requirement for in-flight oxygen. FEV1 and SpO2 have been demonstrated to be poor predictors of desaturation with no established reliable upper limit of PaO2, above which patients will not desaturate. We investigated whether there were lower-limit thresholds, beyond which oxygen is always required and hence HCT can be avoided.

Methods Retrospective analysis of all hypoxic challenge tests conducted at our centre between 2010 and 2017 was undertaken. Baseline demographics, diagnosis and contemporaneous lung function data was recorded. HCT was performed as per BTS guidance and included baseline resting blood gas followed by a repeat after 20 min inspiring 15% FiO2. If PaO2 was <6.65 kPa or SpO2 <85%, 2 L oxygen via nasal cannulae was applied and a repeat blood gas performed to confirm PaO2 ≥6.65 kPa.

Results HCT was performed on 170 occasions during the study period. COPD was the underlying diagnosis in 110 (64.7%) of tests, ILD in 40 (23.5%) and CF (13, 11.8%). Average age (median [range]) was 67.5 years [49.1–83.8]COPD, 67 [52.3–83.3]ILD, 32.5 [19.1–66.8]CF. Lung function (FEV1%pred) was 49.7[21–115]COPD, 71.6 [31–124]ILD, 36.5[23–65]CF. Following HCT, in-flight oxygen was recommended in 99 (58.2%) patients all of whom were recommended 2 l/min. A threshold of <7.55 Kpa on resting blood gas was 100% predictive for requirement of in-flight oxygen and a threshold of <8 kPa was 97.9% predictive. Incorporating the <7.55 kPa and <8 kPa thresholds into clinical practice by proceeding straight to 2 l oxygen could negate the need for HCT in 20.6% and 43.9% of cases respectively.

Conclusion HCT is a useful tool for assessing the need for in-flight oxygen in lung diseases but is a resource heavy test and requires multiple blood samples taken from patients. Our data suggests that there are lower-limit thresholds for resting PaO2 beyond which HCT can be avoided in a significant proportion of patients.