BACKGROUND: Prescription of an appropriate exercise training intensity is critical to optimise the outcomes of pulmonary rehabilitation; however, prescribing cycle ergometry training is challenging if peak work is unknown. Recently two studies reported regression equations which allow estimation of peak cycle work rate from the 6-minute walk distance (6MWD) in chronic obstructive pulmonary disease (COPD). OBJECTIVES: To compare estimates of peak work and target training work rate (60% peak) obtained from these equations. METHODS: Sixty-four (38 male) subjects, mean ± SD age 70 ± 8 years and FEV(1) 49 ± 18% predicted with COPD performed the 6-minute walk test according to a standardised protocol. Estimates of peak work were obtained using the published equations and agreement was examined using Bland and Altman plots. RESULTS: Mean 6MWD was 376 ± 86 m compared to 464 ± 110 m and 501 ± 83 m in samples used to derive the equations. There was substantial variation in estimates of peak work between equations (range 1-75 Watts difference) with a coefficient of variation of 35%. Differences were greater in men than in women (p < 0.001). The Luxton equation predicted higher peak work than the Hill equation in younger subjects and at work rates over 50 Watts. Estimated training work rate differed by more than 20 Watts in 18 subjects (28%). CONCLUSIONS: This comparison of reference equations for predicting peak cycle work rate from 6MWD indicates substantial variation between methods that differs systematically across the range of work rates. Further research is required to validate the equations and assess their utility for exercise prescription in pulmonary rehabilitation.