The aim of this study was to evaluate the influence of different apparatuses, procedures and calculations on the precision of the optimized carbon monoxide method. Total haemoglobin mass was determined twice on consecutive days in 11 subjects using both venous and capillary blood samples. To estimate loss of carbon monoxide due to exhalation, carbon monoxide concentration was measured by two portable carbon monoxide analysers (Fluke CO-220, Fluke, Norwich, UK and Pac 7000 Carbon Monoxide, Draeger Safety, Northumberland, UK) and alveolar ventilation was specified using an automated metabolic gas analysis system (Cosmed Quark b2, Cosmed, Rome, Italy). Blood volume was derived from total haemoglobin mass using haemoglobin concentration and haematocrit obtained in both the supine and the seated position. Two different formulae to calculate blood volume were also compared. Precision was good for both total haemoglobin mass and blood volume measurements performed on consecutive days (typical error < 2%). Using Fluke CO-220 analyser, an estimated alveolar ventilation and capillary blood, total haemoglobin mass (917 +/- 136 g) was similar when compared to the Pac 7000 Draeger CO-analyser (904 +/- 137 g; mean bias 13 g with 95% limits of agreement 26 to + 1 g, P=0.76), specified alveolar ventilation (911 +/- 132 g, mean bias 6 g with 95% limits of agreement 18 g to + 6 g, P =0.87) and venous blood (917 +/- 134 g, mean bias 0 g with 95% limits of agreement 38 to + 38 g, P=0.99), respectively. Blood volume determination was also not significantly affected by the supine vs. seated position (6.7 +/- 0.8 l vs. 6.6 +/- 0.8 l, P=0.56) but can deviate by 0.6 l (P=0.01) depending on the formula applied. Thus, the good precision of the assessment of total haemoglobin mass and blood volume using the optimized carbon monoxide rebreathing method is not significantly influenced by the make of CO analyser, method of obtaining alveolar ventilation, blood sampling method and subject position, but should for longitudinal monitoring purposes use the same formula.