Introduction The benefit of ambulatory oxygen (AO) in patients with hypoxaemia on exercise remains contentious. Often AO is not used in the way the prescribers envisage. Device suitability or poor understanding about the benefits of AO may mean patients remain hypoxic on activity and/or avoid exercise.

Objective To assess patients activity and preference using Homefill^TM versus usual AO device in a mixed population of patients with exercise hypoxaemia and/or LTOT. Homefill^TM allows patient refill of oxygen cylinders (1.9 L) as needed using a compressor.

Methods Inclusion criteria included current use of AO and stable physical condition. AO prescription was optimised for each device. Patients were randomised to usual AO then HomefillTM for 6 wks or vice versa. Tri-axial activity monitors were used during the last week. Patient preference was identified by questionnaire. Weekly calls encouraging activity and AO use were made. If patients suffered an inter-current illness, the trial arm was extended by 2 wks; those with >1 illness were withdrawn. Power calculation indicated a sample of 40 subjects was required to detect a difference of 1,000 domestic activity counts at a 5% significance level with 80% power.

Results 70 patients met the inclusion criteria and 40 enrolled. Mean age 66 yrs, 17 males, 70% had COPD with median FEV1 41% predicted (range 27–71%), restrictive median FVC 70% predicted. 29 complete data sets were collected. There were 9 episodes of AECOPD. Fifteen patients required the highest 02 pulsed setting on HomefillTM There was no statistically significant difference in mean daily activity counts when using HomefillTM compared to usual AO. A decline in activity counts was observed in both cohorts during the second period. Eighteen patients elected to keep HomefillTM of whom 11 previously used LOX as their usual AO.

Conclusions Homefill was equivalent to usual provision of AO and was preferred by the majority. Disappointingly, regular phone encouragement failed to increase activity levels. Activity levels were very low and highly variable reflecting advanced disease/deconditioning.

Implications for Practise Patient use and preference of AO device includes non physiological aspects. AO maybe best targeted at patients before exercise tolerance is severely limited.