INTRODUCTION: Impaired physical fitness may contribute to functional limitations and disability after stroke. Physical fitness (including cardiorespiratory fitness and muscle strength/power) can be improved by appropriate fitness training; this is of benefit to healthy people and patient groups but whether it is of benefit for people after stroke is unclear. The aim of this thesis was to determine whether physical fitness training is beneficial after stroke. OBJECTIVES: (1) Develop a rationale for fitness training by determining whether physical fitness after stroke is a) impaired, and b) associated with functional limitations and disability. (2) Develop and evaluate randomized controlled trial (RCT) evidence by a) determining the feasibility of a definitive RCT, and b) evaluating the benefits of fitness training after stroke. METHODS: (1) Systematic review of observational data and multiple linear regression of exploratory RCT baseline data determined the nature of fitness impairments and any associations with functional limitation and disability. (2) Systematic review and meta-analysis of RCTs established the effects of fitness training on disability, death and dependence. An exploratory RCT (‘STARTER’) compared the effects of a fitness training programme (cardiorespiratory plus strength training 3 days/week for 12 weeks) with an attention control (relaxation) on fitness, function, disability, mood and quality of life in 66 ambulatory people with stroke. RESULTS: (1) Systematic review of observational data showed cardiorespiratory fitness (peak oxygen uptake and economy of walking) and muscle strength were low after stroke; the impairments predicted functional limitation but links to disability were unclear. STARTER baseline data showed little impairment in economy of walking but lower limb extensor power was impaired (42-54% of values expected in healthy age and gender matched people) and this predicted functional limitation and disability. (2) The systematic review identified 12 RCTs (n=289) in 2003, and 24 RCTs (n=1147) when updated in 2007. The systematic reviews showed death was uncommon, and effects on dependence and disability were unclear. However training did improve fitness and cardiorespiratory training during rehabilitation improved ambulation. Most benefits resulted from task-related training. The STARTER fitness training intervention was feasible, with good attendance (>90%) and good compliance with intervention content (94-99%). At the end of the fitness training intervention there were small improvements in some cardiorespiratory fitness, physical function and quality of life outcomes compared with the control group, but these differences had diminished four months later. CONCLUSIONS: (1) Cardiorespiratory fitness, muscle strength and power are impaired after stroke, so there is scope to increase fitness, and there are plausible benefits. (2) Physical fitness training after stroke is feasible, it improves fitness and has some functional benefits, in particular for walking ability. Effects on disability, death and dependence are not known. Further research is required to determine the timing, mode, duration, frequency and intensity of fitness training for optimum benefits, and investigate how benefits can be retained in the long-term.