Background: To determine the longitudinal association of poststroke fatigue with activities of daily living (ADL), instrumental ADL (IADL) and perceived health-related quality of life (HRQoL) and to establish whether this relationship is confounded by other determinants. Methods: A prospective cohort study of stroke patients consecutively admitted for inpatient rehabilitation was conducted. ADL, IADL and HRQoL were assessed in 223 patients at 6, 12 and 36 months after stroke. Fatigue was determined by the Fatigue Severity Scale. Random coefficient analysis was used to analyze the impact of fatigue on ADL, IADL and HRQoL. The association between fatigue and outcome was corrected for potential confounders, i.e. age, gender, comorbidity, executive function, severity of paresis and depression. The covariate was considered to be a confounder if the regression coefficient of fatigue on outcome changed by >15%. Results: Fatigue was significantly related to IADL and HRQoL but not to ADL. The relation between fatigue and IADL was confounded by depression and motor impairment. Depression biased the relation between fatigue and HRQoL, but fatigue remained independently related to HRQoL. Conclusions: Fatigue is longitudinally spuriously associated with IADL and independently with HRQoL. These findings suggest that in examining the impact of poststroke fatigue on outcome, one should control for confounders such as depression.