Abstract Background Early rehabilitation can reduce ventilation duration and improve functional outcomes in critically ill patients. Upper limb strength is associated with ventilator weaning. Passive muscle loading may preserve muscle fibre function, help recover peripheral muscle strength and improve longer term, post-hospital discharge function capacity. The physiological effects of initiating rehabilitation soon after physiological stabilisation of these patients can be concerning for clinicians. This study investigated the feasibility of measuring metabolic demand and the safety and feasibility of early upper limb passive ergometry. An additional comparison of results, achieved from simultaneous application of the methods, is reported. Methods This was an observational feasibility study undertaken in an acute teaching hospital’s General Intensive Care Unit in the United Kingdom. Twelve haemodynamically stable, mechanically ventilated patients underwent 30 minutes of arm ergometry. Cardiovascular and respiratory parameters were monitored. A Friedman test identified changes in physiological parameters. A metabolic cart was attached to the ventilator to measure oxygen uptake. Oxygen uptake was concurrently calculated by the reverse Fick method, utilising cardiac output from the LiDCO™ and paired mixed venous and arterial samples. A comparison of the two methods was made. Data collection began 10 minutes before ergometry and continued to recovery. Paired mixed venous and arterial samples were taken every 10 minutes. Results Twelve patients were studied; 9 male, median age 55 years, range (27–82), median APACHE score 18.5, range (7–31), median fraction inspired oxygen 42.5%, range (28–60). Eight patients were receiving noradrenaline. Mean dose was 0.07 mcg/kg/min, range (0.01–0.15). Early ergometry was well tolerated. There were no clinically significant changes in respiratory, haemodynamic or metabolic variables pre ergometry to end recovery. There was no significant difference between the two methods of calculating VO₂ (p = 0.70). Conclusions We report the feasibility of using the reverse Fick method and indirect calorimetry to measure metabolic demand during early physical rehabilitation of critically ill patients. More research is needed to ascertain the most reliable method. Minimal change in metabolic demand supports the safety and feasibility of upper limb ergometry. These results will inform future study designs for further research into exercise response in critically ill patients. Trial Registration Clinicaltrials.gov No. NCT04383171. Registered on 06 May 2020 - Retrospectively registered. http://www.clinicaltrials.gov.